Skip to main content

Full text of "Year book - Carnegie Institution of Washington"

See other formats

mmmm'.- -m 


^^e^r yy?< 

Digitized by the Internet Archive 

in 2012 with funding from 

LYRASIS Members and Sloan Foundation 


HjOKj Oodc 58 

JULY 1, 1958-JUNE 30, 1959 




Library of Congress Catalog Card No. 3-16716 







Mount Wilson and Palomar Observatories 41 

Geophysical Laboratory 79 

Department of Terrestrial Magnetism 223 

Joint Committee on Image Tubes for Telescopes 307 

Department of Plant Biology 313 

Department of Embryology 361 

Department of Genetics 425 



Report of the Executive Committee 463 

Report of Auditors 465 
Abstract of Minutes of the Sixty-first Meeting of the Board of Trustees 481 



INDEX 493 


Caryl P. Haskins 

Walter S. Gifford, Chairman 

Barklie McKee Henry, Vice-Chairman 

Robert Woods Bliss, Secretary 

James F. Bell 
Robert Woods Bliss 
Amory H. Bradford 
Omar N. Bradley 
Vannevar Bush 
Walter S. GifTord 
Crawford H. Greenewalt 
Caryl P. Haskins 
Barklie McKee Henry 
Alfred L. Loomis 
Robert A. Lovett 
Keith S. McHugh 
Margaret Carnegie Miller 
Henry S. Morgan 
Seeley G. Mudd 
William I. Myers 
Richard S. Perkins 
Henning W. Prentis, Jr. 1 
Elihu Root, Jr. 
Henry R. Shepley 
Charles P. Taft 
Juan T. Trippe 
James N. White 
Robert E. Wilson 

1 Died October 29, 1959. 

TRUSTEES Continued 


Barklie McKee Henry, 
Robert Woods Bliss 
Walter S. Giflford 
Caryl P. Haskins 
Robert A. Lovett 


Henry S. Morgan 
Henning W. Prentis, Jr. 
Henry R. Shepley 
James N. White 


James N. White, Chairman 
Walter S. Giflord 
Alfred L. Loomis 
Henry S. Morgan 
Richard S. Perkins 
Henning W. Prentis, Jr. 


Keith S. McHugh, Chairman 
Alfred L. Loomis 
Juan T. Trippe 


Henry S. Morgan, Chairman 
Robert Woods Bliss 
Walter S. Gitford 
Barklie McKee Henry 


Omar N. Bradley, Chairman 
Barklie McKee Henry 
Henry S. Morgan 
James N. White 


Seeley G. Mudd, Chairman 
Crawford H. Greenewalt 
Elihu Root, Jr. 


Alfred L. Loomis, Chairman 
Margaret Carnegie Miller 
William I. Myers 
Charles P. Taft 


Juan T. Trippe, Chairman 
Barklie McKee Henry 
Henning W. Prentis, Jr. 
Robert E. Wilson 


Henry R. Shepley, Chairman 
James F. Bell 
Robert Woods Bliss 
Juan T. Trippe 

Note: Membership of Committees as of June 30, 1959. 




Daniel Coit Gilman, 1902-1904 Robert Simpson Woodward, 1904-1920 

John Campbell Merriam, President 1921-1938; President Emeritus 1939-1945 
Vannevar Bush, 1939-1955 

Alexander Agassiz 
George J. Baldwin 
Thomas Barbour 
John S. Billings 
Lindsay Bradford 
Robert S. Brookings 
John L. Cadwalader 
William W. Campbell 
John J. Carty 
Whitefoord R. Cole 
Frederic A. Delano 
Cleveland H. Dodge 
William E. Dodge 
Charles P. Fenner 
Homer L. Ferguson 
Simon Flexner 
W. Cameron Forbes 
James Forrestal 
William N. Frew 
Lyman J. Gage 
Cass Gilbert 
Frederick H. Gillett 
Daniel C. Gilman 
John Hay 
Myron T. Herrick 
Abram S. Hewitt 
Henry L. Higginson 
Ethan A. Hitchcock 
Henry Hitchcock 
Herbert Hoover 
William Wirt Howe 
Charles L. Hutchinson 
Walter A. Jessup 
Frank B. Jewett 
Samuel P. Langley 
Ernest O. Lawrence 
Charles A. Lindbergh 
William Lindsay 
Henry Cabot Lodge 

Under the original charter, from the date of organization until April 28, 1904, the following were 
ex officio members of the Board of Trustees: the President of the United States, the President of the Senate, 
the Speaker of the House of Representatives, the Secretary of the Smithsonian Institution, and the President 
of the National Academy of Sciences. 



Seth Low 



Wayne MacVeagh 



Andrew W. Mellon 



Roswell Miller 



Darius O. Mills 



S. Weir Mitchell 



Andrew J. Montague 



William W. Morrow 



William Church Osborn 



James Parmelee 



Wm. Barclay Parsons 



Stewart Paton 



George W. Pepper 



John J. Pershing 



Henry S. Pritchett 



Gordon S. Rentschler 



David Rockefeller 



Elihu Root 



Julius Rosenwald 



Martin A. Ryerson 



Theobald Smith 



John C. Spooner 



William Benson Storey 



Richard P. Strong 



William H. Taft 



William S. Thayer 



James W. Wadsworth 



Charles D. Walcott 



Frederic C. Walcott 



Henry P. Walcott 



Lewis H. Weed 



William H. Welch 



Andrew D. White 



Edward D. White 



Henry White 



George W. Wickersham 



Robert S. Woodward 



Carroll D. Wright 





813 Santa Barbara Street, Pasadena 4, California 
Ira S. Bowen, Director; Horace W. Babcock, Assistant Director 

Halton C. Arp 
William A. Baum 
Armin J. Deutsch 
Jesse L. Greenstein 

Fred Hoyle 

Rudolph L. Minkowski 
Guido Munch 
J. Beverley Oke 

Allan R. Sandage 
Olin C. Wilson 
Fritz Zwicky 

5241 Broad Branch Road, N. W ., Washington 15, D. C. 

L. Thomas Aldrich 
Ellis T. Bolton 
Roy J. Britten 
Bernard F. Burke 

Merle A. Tuve, Director 

Dean B. Cowie 
John W. Firor 
Scott E. Forbush 
W. Kent Ford, Jr. 
Norman P. Heydenburg 

Richard B. Roberts 
Georges M. Temmer 
Harry W. Wells 
George W. Wetherill 


2801 Upton Street, N. W ., Washington 8, D. C. 

Francis R. Boyd, Jr. 
Felix Chayes 
Sydney P. Clark, Jr. 
Gordon L. Davis 

Philip H. Abelson, Director 

Gabrielle Donnay 
Joseph L. England 
Hans P. Eugster x 
Joseph W. Greig 
Thomas C. Hoering 2 

1 Resigned September 30, 1958. 

2 Appointed January 1, 1959. 

Gunnar Kullerud 
J. Frank Schairer 
George R. Tilton 
Hatten S. Yoder, Jr. 

STAFF Continued 


Stanford, California 
C. Stacy French, Director 

William M. Hiesey Malcolm A. Nobs 

Harold W. Milner James H. C. Smith 


Wolfe and Madison Streets, Baltimore 5, Maryland 

James D. Ebert, Director 

David W. Bishop Elizabeth M. Ramsey 

Bent G. Boving Mary E. Rawles 

Robert K. Burns Royal F. Ruth 
Robert L. DeHaan 


Cold Spring Harbor, Long Island, New Yor\ 

Milislav Demerec, Director 

Alfred D. Hershey Margaret R. McDonald 

Berwind P. Kaufmann George Streisinger 

Barbara McClintock 

STAFF Continued 


1530 P Street, N. W ., Washington 5, D. C. 

Caryl P. Haskins 

Paul A. Scherer 

Executive Officer 

Edward A. Ackerman 

Deputy Executive Officer 

Samuel Callaway x 

Assistant to the President 

Ruth L. McCollum 

Assistant to the President 

Ailene J. Bauer 

Director of Publications 

Lucile B. Stryker 

Earle B. Biesecker 

Bursar; Secretary-Treasurer, Retirement Trust 

James W. Boise 

Assistant Bursar; Assistant Treasurer, Retirement Trust 

James F. Sullivan 

Assistant to the Bursar 

Richard F. F. Nichols 

Executive Secretary to the Finance Committee 

1 Retired June 30, 1959. 

STAFF Continued 


of Carnegie Institution of Washington 

D. G. Catcheside, University of Birmingham 
Georges N. Cohen, Institut Pasteur 

Hessel de Vries, University of Groningen 

Louis B. Flexner, University of Pennsylvania 

Willard F. Libby, University of California at Los Angeles 

Kenneth McQuillen, Cambridge University 

Paul W. Merrill, Mount Wilson Observatory 

Jan Hendrick Oort, Leiden Observatory 

Harry E. D. Pollock, Carnegie Institution of Washington 

E. E. Salpeter, Cornell University 

Evelyn E. B. Smith, University of Glasgow 

P. Swings, Universite de Liege 

C. E. Tilley, Cambridge University 

M. Westergaard, Universitetets Genetiske Institut, Copenhagen 

Evelyn M. Witkin, State University of New York 

R. v. d. Woolley, Royal Greenwich Observatory 


All over the world, contaminated by our immediate past errors, we strive to 
bring bac\ order, scruple, and principle into societies, which willy-nilly face 
a crowding, narrowing, common future .... Everywhere everyone seeks a 
new focus to lend form, purpose and order to the citizen's life. Whence is it to 

The Times Literary Supplement, July 17, 1959 

In the society of scientists each man, by the process of exploring for the truth, 
has earned a dignity more profound than his doctrine. A true society is sus- 
tained by the sense of human dignity .... This is why, at bottom, the society 
of scientists is more important than their discoveries. What science has to teach 
us here is not its techniques but its spirit: the irresistible need to explore. 

J. Bronowsk} — Science and Human Values 

profound change and radical adjustment, such as those in which we 
live, permit and indeed demand far more than simple endurance, or 
talent in survival. Such times are fraught with perils, as we are acutely aware. 
Yet hazard is not their only quality, or even their dominant one. They offer 
far more than mere tests of ingenuity in living, to be met by essentially con- 
ventional, if strenuous, feats of adaptation and accommodation. Met with 
boldness and audacity, with balance and with dedication, and above all with 
independence of mind and spirit, these may also be times of unmatched oppor- 
tunity for new orders of growth and self-realization, whose ultimate benefits 
could far outweigh the short-term perils that preoccupy and shadow our 

It is interesting to consider how those earlier generations living in times of 
comparable stress and change — many of which we, at this perspective, call 
Golden — may have estimated their lot: the citizens of Periclean Athens, for 
example, or of Rome in the early days of the Republic, or Italy in the late 
fifteenth and early sixteenth centuries, or the Low Countries in Renaissance 
times, or Elizabethan England. All these periods had certain features in 
common. All inherited a background of unusual technical accomplishment, 
upon the power of which, together with power economically and socially 
derived, much of their great opportunity for achievement rested. In each an 
unusual degree of social and political stability had been attained, at least 
briefly, at home. Yet the threat of violence and social disorganization can 
never have been very far away. Men knew that they lived in uncertainty and 
danger. But they also realized the enormous challenge of their days, and one 
can sense in them a prescient eagerness and a new audacity. A flavor of the 
immense significance of individual effort and individual achievement carries 
to our own day. A certain dynamic fluidity — a special ability, perhaps a 
special eagerness, to embrace the condition of change for its own sake, re- 


gardless of its specific content — colors each history. All of them, as Elting 
Morison has pointed out, seem to have attached a higher value to the dynamic 
process of living in their society than to the material consequences — the static 
products of the skill and industry of that society. Clearly, new ideas were 
valued and exciting, and their communication was as important as their genesis. 
Exploration, questing innovation in one or another realm of experience, had 
an especially honored place. 


lO PATTERNS, of course, are ever precisely repeatable, no history 
ever exactly recapitulated. The interlocked character of the modern world, 
the growing external similarities of its cultures, the universality of com- 
munication, the immense complexity of intellectual and social and ma- 
terial environments, all obscure detailed comparison with earlier times. And 
yet there is much that might be compared — the political stability coupled with 
the apprehensions of widespread chaos and violence, the long history of 
technical achievement, the tradition of individual initiative and innovation 
once more flowing like an incoming tide in a world where it has been too 
long in ebb, the preoccupation with modes of organizing human effort, the 
delight in ideas and their communication. 

It is not surprising that we should find such parallels with other ages which, 
in the perspective of the years, we count brilliantly successful. In a general 
sense, of course, both the opportunities and the hazards were the same as those 
we face today — the opportunities measured in the great rewards that foresight 
and understanding and the action implementing them must always bring in a 
rapidly changing world; the hazards, likewise, those that failure of adequate 
reach and grasp have always brought — the grinding penalties of confinement 
to the narrows and the shallows; at worst, the total forfeiture of a people's 
birthright in the world. 

But in certain ways our challenge differs from that confronting earlier 
periods so greatly in degree as perhaps to be of another kind. All the earlier 
ages of greatness, to be sure, ever since crops were deliberately planted or the 
first crossbow designed, were seriously concerned with technical innovation and 
development, and depended upon that resource for much of their superiority. 
But in no earlier society can the challenge to acuity of comprehension, to 
specialized skills, to generalized insight, to the fullest development of individual 
capability in every field, have been remotely so imperative, nor can the 
demand and the opportunity for individual development have approached 
those of our own day. Ours is surely an age in which the stimulus, the demand, 
and the opportunities for innovation, for individual creativeness, have reached 
unprecedented heights. 

A curious qualitative difference, also, inheres in the practical technical 
demands of our times. Every dynamic age has sought hungrily for the sources 
of physical power to shape the natural world to its ends, and much of its 


technical strategy has been directed to that purpose. From the latter half of 
the eighteenth century almost to the present that search perhaps stood first 
among the challenges of technology. It is still an important consideration for 
all industrial nations. But, precisely because the quest has been so successful, 
we may perhaps expect that in the future other aspects of our conquest of 
nature, concerned more pointedly with understanding and control, will assume 
a more conspicuous role in our technology, as is happening even now. Such 
aspects demand a somewhat different practical approach to the natural world, 
posing a challenge on every front to greater subtlety and penetration. In the 
future they must emphasize further and give a yet more keen incentive to a 
broader, and possibly a subtler and deeper, comprehension of nature as a practi- 
cal requirement of living. But there will be other incentives, too — more pro- 
found, and ultimately more imperative and lasting. 


NCE again, for nations as for men, times of unusual stress and challenge 
are a supreme test, not alone of effectiveness in action, but primarily of inner 
cohesion and verve, of inner wholeness. However severe the practical challenge, 
it can be successfully met by a people only in its own characteristic fashion, 
consonant with its own most deeply held values. The whole strength, the 
whole balance, of a nation must rest on how far what it does undergirds those 
values. Whenever a mode of life, however outwardly successful, inwardly leads 
to distrust or denial of them, the radiance of an age of greatness can be quickly 
dimmed. It is instructive, indeed, to see how speedily and how poignantly 
some of the eras of intellectual greatness have been thus affected. 

Surely there has never been a greater period in the history of science than 
that of Newton. The time of his most important work, the mid-seventeenth 
century, has perhaps never been surpassed in the sheer brilliance and variety 
of its intellectual and cultural contributions. These were the years of the 
founding of the Royal Society and the French Academy, and of the work 
of a host of brilliant innovators in fields ranging from science through litera- 
ture and all the arts. Yet only sixty years later the values that motivated the 
age had clearly become infected with disbelief (well exemplified, for instance, 
in much of the poetry of Pope), and, perhaps in consequence, a curious dull- 
ness settled over it. It was as though the essential sources of inspiration were 
no longer reinforced, but instead had become the objects of good-natured 
derision, or even of savage calumny. Before such onslaughts commitments to 
values of this depth and seriousness may not permanently disappear. But they 
are gravely damaged. The creative mood that rests upon them can diminish, 
in nations as in men, with frightening speed. 

We are not immune to similar dangers. We are faced by practical technical 
challenges of a national and international character in which the seeking for 
knowledge and understanding of nature for its own sake is often enough united 
in mutual dependence with material achievements indispensable to our pros- 


perity or even essential to our survival. In such a situation it is all too easy to 
mistake the end of the road for the road itself, and even to be tempted or 
forced into the unwise and unenviable role of the anxious emulator. In these 
days, especially, it is imperative that we be clear about our deepest motivations, 
that we meet these challenges in our own way and by courses that will rein- 
force our own ethic. We must strain every effort to avoid the widespread con- 
fusion of process with product that continually threatens us. It is essential that 
all of us, whether directly concerned or not with scientific matters, comprehend 
the values that underlie the search for scientific verities, and appreciate how 
vividly they mirror our more generalized ethic. 

In the philosophy of our society three anchor points stand out with special 
prominence. The first is that primary dedication to the individual and to his 
creativeness, that overriding commitment which impels us to see the stature 
of our nation primarily as a function of the welfare and perfection of realiza- 
tion of the worth and stature of the individual — perhaps our greatest inherit- 
ance from the Reformation — which it has been our particular national pride 
to cultivate and foster. On it depend the fundamental orientation and integrity 
of our society — this, indeed, must be its primary raison d'etre. 

Closely linked with this value is our basically religious culture. As 
de Tocqueville long ago observed, religion in very fact gave birth to large 
sectors of our society, and still profoundly influences our deepest national atti- 
tudes and motivations. We are accustomed to identify a strong drive to indi- 
vidual accomplishment and achievement with the Calvinist tradition that has 
come from New England cultural and spiritual forebears. Though we identify 
the source too narrowly, we are surely right that a large share of that drive 
has nonmaterialistic roots. 

A third wellspring of our national life is our deep respect for, and concern 
with, both learning and innovation, and perhaps especially the exploration of 
nature. Throughout our history we have traditionally loved the challenge of 
technical change. Even the lone inventor is surely one of the most character- 
istic — if also one of the most uncertainly rewarded — of our historic figures. 

The most powerful indication that the mood of research in science, taken as 
the search for knowledge of the natural world purely for its own sake, is 
deeply indigenous to our culture is the degree to which the pre-eminent values 
and motivations of that mood epitomize and undergird all three of these main- 
springs of our culture. In this, of course, science shares with creative intel- 
lectual activity in every field. Indeed, as Martin Johnson has pointed out, 
science and art may well be fundamentally at one in their deepest objectives 
and concepts as well as in their values — the structuring of knowledge in com- 
municable patterns and forms, whether in the domain of feeling or in that of 
quantitative measurement. But in its dedication of individual effort, its pri- 
mary seeking after universally communicable knowledge of the natural world, 
in the discipline of its economy and its parsimony, in its sense of style, scientific 
research must be reckoned, in our day, one of the vital bulwarks of our deepest 
national values. 


That bulwark serves us in a particularly critical time. Thoughtful European 
observers of our current scene have been quick to remark that the strongest 
motivations for our striving are part of an essentially nonmaterialistic ethic. 
But they have not failed to notice, too, that our very material success poses a 
strong and continuing threat to that ethic; that the understanding of the real 
nature of our deepest motivations may be in danger of blurring, and that what 
American society may need more than anything else today is a more explicit 
image to give us continuing belief and confidence. Science may have an im- 
portant part of that image to provide. 


1 UT there is an obverse to this coin. In a world which now, for better or for 
worse, is basically technical in character, in which technological power, as never 
before, is essential for survival itself, let alone for material prosperity, the 
very phrase "science and technology" carries with it the danger of confusing 
application with creation, the danger that the scientific process may be widely 
thought of primarily as an ever-expanding fount of technical products. Such 
a view carries the very real hazard of revulsion on the part of many not aware 
by direct experience of the nature of the scientific mode, a revulsion partly 
from a realm which is felt to be important but is not understood, but, much 
more significantly, from an activity thought to be essentially materialistic in 

It is somewhat startling to realize that there may indeed be a wide gulf 
between men whose training, concerns, and values have lain in the general 
areas of scientific pursuits, and men of nontechnical background, as Sir Charles 
Snow has recently pointed out. More serious than the gulf of formal intel- 
lectual content — since the closing of this is only a challenge, though yet a 
formidable one, to education and communication — is the accompanying danger 
that the divergence of experience set and harden a real and a fundamental 
dichotomy of attitudes and values between the two worlds, and encourage an 
unbridgeable hostility between them. 

So far as communication of substance is concerned, the case is of course 
quite clear. What separates the groups is a whole nexus of conditioning and 
background. It is as though a people, migrating into an environment com- 
pletely different from any they had ever known, exposed now to concepts en- 
tirely new, were obliged to invent wholly novel terms for wholly novel entities, 
and the new language, interacting with the concepts underlying it, as always 
happens, had in turn bred new thinking, and the whole process, over the years, 
had accelerated and ramified until one group finally was only remotely in- 
telligible to another. 

This in itself is neither a novel nor a necessarily critical situation. Similar, 
if somewhat less formidable, problems of communication obtain within the 
framework of the scientific effort itself. Within the framework of the scientific 
way, there are enormous differences not only in the subject matter and terminol- 


ogy of various disciplines, but in attitudes, modes, states of evolution, degrees 
of maturity. Men of the most diverse temperaments and talents are enlisted 
in the cause of science. So great are their numbers, so diverse their concerns, 
that few, during their entire lives, can even be aware of, let alone critically 
apprehend, each other's work. And yet basically they are of one group, since 
they share the most fundamental things — the motivation, the philosophy, the 
total orientation of the scientific way. It is clear that the threat of a fundamen- 
tal dichotomy between the scientific and nonscientific worlds almost certainly 
rests, primarily, not so much on unfamiliarity of subject matter as on a 
supposed irremediable divergence or misunderstanding of values. 

This sense of division between scientific and other creative effort is inherently 
dangerous. Because of the enormous volume of scientific work and its com- 
manding importance in our society, such a dichotomy, if based in a misread- 
ing of values which is unavoidable, could involve a number of serious hazards. 
Consider one of the most striking — the destructive social competition inherent 
in the enlisting of young minds in the scientific way which would be an in- 
evitable consequence. 

The total expenditures formally catalogued for research and development in 
the nation for 1958, according to some calculations, exceeded by a factor of 
more than three those similarly catalogued for 1950. For the decade of the 
1950's it has been estimated that the total will have been of the order of about 
sixty billion dollars, and it has been forecast that for the decade of the 1960's 
it may reach one hundred and sixty billion. These are startling figures, and 
of course only a fraction accounts for scientific research in its fundamental, 
uncommitted aspect. Yet, even allowing for this, there is no gainsaying the 
sharp upward trend in the volume of effort involved in some fashion with 
science in our day, and in the proportion of the members of our society directly 
concerned in it. This trend is by no means new. In England there is good 
indication that since the time of Newton the trend of the total work con- 
nected with science has been rather consistently exponential, effecting a 
doubling of the volume of the scientific effort approximately every ten to 
fifteen years — three times in a generation. It seems reasonable to assume that 
this rate is likely to continue for some time. Moreover, it seems to be ap- 
preciably greater for scientific endeavor than for many other activities. 

This situation carries many and fascinating implications. The range and 
sweep of the scientific effort being what they are, it would appear that the sci- 
entific and technical mode is not only remaking our culture materially. Super- 
ficially, it seems that it must, ultimately, push to the wall those other modes 
by which we have lived in the past and which remain so precious to us: art 
and literature and humanism, and all those creative activities of a nonscientific 
nature. With science growing exponentially at such a rate, will it not simply 
take over all the excellent young minds, leaving none eventually for the other 
departments of thought, and eventually exhausting the whole supply? 

If the gulf separating the realms of science and other departments of human 
activity is indeed unbridgeable, involving irreconcilable differences of value and 


motivation, then such a reflection is disturbing indeed. For it follows that 
science must inevitably be permanently competitive with other areas of human 
effort. This is not a welcome prospect — most of all because it could reduce the 
multivariant range of skills and interests in our society which must, rather, 
be encouraged and expanded if that society is to remain at its most versatile 
and vigorous. 

But there is heartening evidence on the other side. It is of course a matter 
of common observation that, perhaps excluding intellects in the category of 
genius and those with gifted qualities of a highly specialized kind, good minds 
in science tend also to be natively good minds in other fields, and often to 
seek spontaneously the common denominators among them. Historically, 
moreover, older periods of cultural burgeoning were also periods of burgeoning 
of early science. Even so late as the early days of the Royal Society it would 
have been hard to characterize many of its members as great scientists or great 
humanists, or both. Furthermore, we are prone to forget that some of the 
most impressive of the scientific disciplines of our own day — some of those we 
properly consider among the most rigorously developed examples of scientific 
concept and method — can be historically traced in an almost unbroken evolu- 
tion of ideas from their present-day rigor and profundity to representations of 
beauty and order in nature on the part of eighteenth-century observers in a 
pattern that we would surely characterize today as more artistic than scientific. 
Indeed, there is much evidence that this has been a more general historic mode 
of origin of the sciences than we commonly realize. The science of genetics, 
for example, offers a vivid and beautiful illustration. 

More penetrating reflections on the same question may be suggested by the 
role played by certain concepts that are central to the structure of post- 
Newtonian science itself, but that prove relevant to ever-widening sectors of 
what we have often conventionally considered the nonscientific world, which 
Bronowski has especially illumined. 

For close on three hundred years the concepts of order and of causality gave 
force and depth, and above all unity, to post-Newtonian science. Perhaps the 
most important task of this science was to put order into experience. Its con- 
cepts rested importantly on two basic assumptions about the world of observa- 
tion. First, that world can be represented as built of units forming combina- 
tions subject to straightforward analysis, units both essentially like in kind, 
whatever their magnitude, and, by implication, largely invariant in time and 
independent of the observer. Second, the relationships between these units 
and the behavior of the models of the real world which they form are governed 
by laws much like those of Euclid: they are precise and can be formulated; 
they can be attained by processes of formal logic; and they will match the con- 
dition of the actual world. The world is thus essentially determinate, es- 
sentially predictable. 

These concepts did much to give post-Newtonian science the intense ana- 
lytical acumen and the formidable power that it displayed in so many of its 
facets. But they also adapted it primarily to a relatively small sector of the total 


world: the macroscopic physical world. Not the microscopic physical world; 
not the world of life; not the world of collectively organized living things. 

Post-Newtonian scientists of this discipline knew quite well — or soon dis- 
covered — that they must restrict this framework of scientific investigation to a 
limited segment of the world if their methods were to be successfully applicable, 
and they ultimately paid with a highly disciplined parochialism for added 
depth and penetration. It was surely true for three hundred years that the 
province of the highly organized sector of the natural sciences concerned with 
rigorous analysis and the ordering of systems did have to reject as not amenable 
to investigation by its methods many broader concerns, and accordingly — 
and necessarily — excluded them from its general purview. To a considerable 
extent this is still true, and this imperative and inherent quality has helped to 
give emphasis over the years to the notion of an unbridgeable dichotomy be- 
tween scientific and extrascientific orientation and concerns. 

But a dramatic widening of the areas of scientific concern, with a progressive 
lowering of old barriers, surely constitutes one of the major revolutions — among 
many — through which we are living today. Central in it is the role of the 
concept of chance and the recognition of its extraordinary significance, which, 
as Bronowski has pointed out, may ultimately prove as great a unifying princi- 
ple for the diverse branches of science as the concepts of cause and effect in the 
days of Descartes and Hobbes. It permits access to a far wider field of experi- 
ence in nature than could be approached earlier by the path of analytical science, 
and it may well be fusing some sectors of what were once quite distinctly scien- 
tific concerns with others earlier considered primarily extrascientific. No more 
vivid example of this, perhaps, can be offered than the quantitative and statisti- 
cal studies of population genetics and population structure that in recent years 
have so illumined our understanding of assemblages of living things. Again, 
the concepts of relativity have accustomed us to the notion that, in any observa- 
tion, the place of the observer in the world that he sees is vital, and that the real 
meaning of truth may be consensus among observers who seek it from different 
standpoints and by different methods. Such concepts bring the mind as a 
living system into as critical focus for the scientific mode as it long has been 
for the humanistic one, setting both modes in a coincidence of concern. 

Living systems, indeed, may today be introducing us to the next — and 
probably very radical — expansion in the ever-widening sector of our world that 
can be brought under scientific cognizance. Since the days of Karl Pearson, 
or even of Sir Francis Galton, we have been acutely aware of the importance of 
the concept of statistical probability, in the living no less than in the nonliving 
world. But when we deal with life we are dealing with self-organizing and 
goal-seeking entities. In a living system, everything happening in that system 
alters the likelihood as to which of a number of possible occurrences will next 
take place, so that the whole series of events making up the past history and 
future prospects of the organism is in fact a closely woven nexus, linking its 
past history and its future prospects in the most complex fashion. The radical 
quality of the idea can be graphically illustrated by posing the old question 


involving the mechanisms of evolution: mutation, selection, survival. Many 
fine scientists of another, but still post-Darwinian, generation, when confronted 
with the question: How could a microcosm, over the aeons, by mutation and 
selection alone, have evolved into a Shakespeare? would have answered: Re- 
gardless of how many generations one conceivably has to work with, regard- 
less of how rigid selection may have been, regardless of how many races of 
organisms may have been extinguished in the advance, regardless of the con- 
ceivable time span involved, it is still not possible to imagine how, by a suc- 
cession of random changes and episodic selection alone, such a development 
could have occurred. The time may be coming, however, when this will not 
appear a meaningful approach. Rather, one may reason: at no stage need it 
be supposed that the process was wholly random in a conventional sense. Each 
evolutionary event within, as well as without, the organism must have biased 
the significance of the next event, and these consequences in turn must have 
been built into a complex forward-moving chain. By the same token, the 
farther the system has moved along this chain, the more elaborate becomes the 
network of complexities to which it is committed, and, the more elaborate the 
history of its development becomes, the slimmer is the chance that it can ever 
be exactly reversed. 

Concepts of this nature, and the research involving conditional probabilities 
that underlies them, are somewhat crude and tentative at present. Yet they 
are surely well past the point of universal skepticism. And considerations of 
this kind, be it noted, are directed at a very large segment of the world that we 
know. The dynamic goal-directed, self-organizing systems to which they may 
apply include the viruses and the single cells. They include the world of many- 
celled plants and animals, and the body and the brain of man. They also in- 
clude populations and societies of living things — and, who knows, perchance 
one day of man himself. 

So, one by one, over the years, the barriers of attitude, of ways of thinking 
and modes of attack which since the time of Newton have traditionally segre- 
gated the concerns of science, have put them in contrast with, and have set them 
apart as narrower and somewhat different from, those of the rest of the world, 
are, in the larger philosophic sense, being penetrated. It seems quite conceivable 
that, as the years go by, our notions about the separateness of the concerns and 
the domain of science from the rest of the world may further blur and fade. 

When and if that occurs, it will come about, not by one view of the world 
replacing or setting aside another, but rather by a convergence on common 
interests and a common kingdom. Such a development may prove to be the 
most powerful of all the levelers of barriers, the most effective of any of the 
bridges spanning the gulf between the scientific and nonscientific viewpoints in 
the years to come. When that happens, perhaps our concern that the exponen- 
tial growth of the scientific effort disastrously pre-empt our human resources 
of mind and spirit will seem to have become largely irrelevant. It may be that 
we will merely experience what mankind has so often experienced before in its 
history. Though our universe will have immensely broadened and deepened 


and to some extent changed its character, as our methods of exploring it have 
changed, our basic view of it may grow in its essential unity. 


T IS clear that the task of uncommitted scientific research is even broader 
and deeper than the winning of new knowledge. It has a banner to maintain, 
within our national ethos, as one of the callings which, among all our manifold 
activities, as a way of life most clearly epitomizes our basic values as a people. 
Its opportunities, its breadth of horizon, its range of concepts and of problems, 
surely exceed, today, those of any other age, not excluding earlier periods which 
we have considered of critical significance to all subsequent scientific develop- 
ment. Within diat framework it has, together with its opportunities, an im- 
portant general obligation to discharge. 

Not only the threat of success, but the very threat of bigness, in a vast techno- 
logical society like our own, must menace the sanctity and maintained sense 
of worth of the individual. In such a society as ours, it is obviously a condition 
of very survival that the work of myriads of individuals be skillfully and suc- 
cessfully marshaled into organized group achievement. Without our aptitude 
in this, one of the most vital elements of our national strength would be lost, 
to our intense detriment and serious danger. But it is far too easy to identify 
the organized group with the members who compose it. As Joyce Cary has 
observed, freedom and independence of mind must involve solitude in thought. 
In sympathy and in feeling we are communal, but in mind each of us is very 
much alone. It is far too easy to forget that the very essence of creativeness 
must lie with the individual as contrasted with the group, whose greatest 
strength may be, at another time and in another place, to develop the fullest 
consequences of his vision. 

This, above all, is what the research environment must stand for. Its tasks 
are many: to provide for the mobility and the freedom of choice and action, 
to furnish the dynamic facilities for broad contacts and wide communication 
and simultaneously the shielding and the opportunity for isolation that may 
be so important to innovation, and indeed so precious to it at its most sensitive 
and vulnerable point; to contribute toward the maintenance of flexibility and 
growth in the disciplines with which it is concerned — helping to guard them 
from a premature and too-rigid ordering of their content and from the insidious 
hardening of dogma. But it has another duty of yet wider and more profound 
significance: to make manifest its spirit as a society devoted to and sustained 
by the earned dignity of the individuals composing it. This is the image of our 
larger society which, as that society must constantly be reminded, alone can 
give it meaning, alone can give it autonomy and authority. Every element of 
the research society, every element of its structuring and adaptation, must con- 
tribute to that end. 

The Year's Work in Review 

As the program of the Institution continues to increase in volume and scope 
and diversity, it becomes continually more difficult to select examples of the 
year's research program for this review. More than ever it must be emphasized 
that the programs described cannot be considered necessarily either more or 
less fundamentally important than others that might have been included. 
They appear here for many reasons: their representativeness of the various 
fields of work in which the Institution is especially occupied at this time; the 
dramatic or decisive character of some of them; even, sometimes, as illustra- 
tions of the failures or the false starts which are as typical of scientific research 
as of other kinds of intellectual adventure. The reader is urged to refer to the 
more extended reviews of the work of the separate Departments for more 
detailed and complete descriptions of work in which he is especially interested. 

The Annual Report for last year summarized the long and distinguished 
pioneering program of the Department of Archaeology and announced the 
completion of the work and the termination of the Department. This does 
not mean, however, that the work of the Institution in this field is ended. On 
the contrary, special phases of it continue to be carried forward most actively 
through a number of research programs, notably those of Pollock and Pros- 
kouriakofr, Eric Thompson, and Shepard. Many special features of the civili- 
zation of the Maya continue to be illumined by these investigations within the 

The work of the Institution continues to be concerned with research in the 
physical and the life sciences, deployed over a very wide front in the Geo- 
physical Laboratory, the Department of Terrestrial Magnetism, the Mount 
Wilson and Palomar Observatories, the Department of Plant Biology, the 
Department of Embryology, and the Department of Genetics. 

The interest of the Institution in the earth sciences is as long-standing as 
the experimental programs of its Departments. The field is a vast one, embrac- 
ing a concern not only with the chemical constitution of the crustal surface of 
the earth and its atmosphere and hydrosphere, now and in the remote past, 
with the nature of its deeper interior, and with the energy changes within the 
system, but also with the receipt of energy from space and the energy losses 



into it, the nature of the electric and magnetic fields surrounding the earth, 
and the phenomena associated with them. The experimental approaches to 
these questions are of the most varied character, including among others the 
techniques of macro- and microchemical analysis, investigations of mineral 
reactions at high temperatures and pressures, studies of X-ray diffraction, and 
techniques of crystallography, of mass spectrometry, of isotope analysis, of 
seismic investigation, and of radio astronomy. 

The Geophysical Laboratory and the Department of Terrestrial Magnetism 
continue to be deeply engaged in many aspects of the earth sciences. A par- 
ticularly important part of the program is concerned with the mechanism of 
formation, the age, and the composition of earth minerals, including especially 
the silicate minerals, which are such a large part of the earth's crust, various 
important ore minerals, and substances of suspected paleobiological origin. 

Work concerned with the major minerals is well illustrated by the studies 
of F. R. Boyd, Jr., J. L. England, H. S. Yoder, Jr., and C. E. Tilley, of the 
Geophysical Laboratory. With special equipment designed to reproduce con- 
ditions of pressure and temperature obtaining within the earth's crust and 
mantle, Boyd and England have explored some mineralogical changes occur- 
ring under pressures up to 50,000 bars, equivalent to a depth of about 150 
kilometers. They have shown that orthopyroxene (a calcium-magnesium-iron 
silicate with manganese) and minerals rich in alumina common in crustal 
rocks will react under conditions of temperature and pressure found in the 
mantle to form the magnesian garnet pyrope. Their data support the hypothe- 
sis that a major mineral in the mantle is pyrope-rich garnet, and that the 
postulated Mohorovicic discontinuity marks a transition from basalt to pyrope- 
containing eclogite (a coarse-grained ferromagnesian silicate rock). These 
garnet-pyroxene mineral assemblages have been considered a potential source 
of basalt, which at least in oceanic areas must come from within the mantle. 
Yoder and Tilley, choosing proper experimental conditions, have converted 
eclogite into basalt and pyroxenite or hornblendite, familiar rocks in the crustal 
layer. By partial melting of eclogite either of two major types of basalt can be 
generated in the laboratory. Thus the hypothesis that eclogites may be a 
potential source of basalt has received some experimental confirmation. 

Studies of diagnostic minerals important in determining conditions of 
metamorphism have included investigations of cordierite, and experimental 
petrological examinations have been made of pyroxenes, garnets, biotites, 
amphiboles, spinels, and feldspars. Zies and Chayes have completed the first 
successful combined micrometric and chemical analysis of a pseudoleucite, 
supplemented by detailed studies of associated groundmass. These studies have 
provided additional information about specific conditions of mineral and rock 
formation and equilibrium situations among related minerals in the earth's 


Studies on the more localized but important ore minerals have been con- 
tinued by G. Kullerud, R. G. Arnold, H. L. Barnes, L. A. Clark, E. H. 
Roseboom, Jr., and R. A. Yund. Investigations of iron-sulfur, copper-sulfur, 
nickel-arsenic, and iron-arsenic binary systems, and iron-sulfur-oxygen and 
iron-zinc-sulfur ternary systems have been completed during the last three 
years ; similar work with nickel-sulfur, iron-arsenic-sulfur, and nickel-arsenic- 
sulfur systems was brought to completion during the current year. Investiga- 
tions of iron-selenium, cobalt-sulfur, iron-nickel-sulfur, copper-iron-sulfur, 
and iron-sulfur-selenium systems are continuing. The simulation of condi- 
tions of temperature and pressure believed to obtain naturally in ore mineral 
formation in these systems continues to provide extraordinarily valuable results. 
These studies offer a most illuminating means for interpreting the relation of 
minerals in a large number of ore deposits found within the earth's crust, and 
provide an excellent tool for investigating the mechanism of ore formation and 
the relation of ore bodies and the surrounding country rock. 

Clues to an understanding of the nature of minerals also are being sought 
at the level of crystal structure. G. Donnay and J. D. H. Donnay at the Geo- 
physical Laboratory, continuing their pioneering investigations, have extended 
the theory of "colored space groups" within crystals as developed by N. V. 
Belov in the Soviet Union, with whom the Donnays have collaborated. They 
have shown that the existence of polar magnetic vectors in a structure intro- 
duces additional symmetry operations which increase the number of space 
groups from 230 to 1421. The application of the extended theory to seven 
mineral compounds of the cubic type of crystal structure was studied during 
the year. 

A program of investigation in the field of isotope geology was initiated this 
year at the Geophysical Laboratory. T. C. Hoering, who joined the staff at 
the beginning of the year, has one mass spectrometer successfully operating 
and another under way. The whole program was speeded by about a year by 
the assistance of Tuve, Aldrich, and Doak, of the Department of Terrestrial 

In continuation of his pioneering studies in paleobiochemistry previously 
reported, Abelson has this year undertaken chemical studies of the highly 
refractory substance kerogen, the most abundant form of all fossil organic 
chemicals. The attack has been directed at two important unsolved questions 
in respect to this very interesting product, undoubtedly originally produced 
biosynthetically but highly modified thereafter: its mode of formation and its 
structure. Abelson's work underlines the possible importance of reactions 
between carbohydrates and peptides or amino acids in the formation of this 
complex constituent of sedimentary rocks. Pilot experiments during the year 
showed that glucose and glycine could be transformed to substances suggesting 
asphaltene or even kerogen, but of different solubility from the natural prod- 


ucts, by incubating mixtures of them at 100 u C and pH values within the range 
that might plausibly have obtained in sedimentary rocks at the time of kerogen 
formation. It may well be that these laboratory reactions represent parts of 
the process by which kerogen is formed under natural conditions. Abelson 
also has split kerogen into petroleum-like substances including hydrocarbons 
through low-temperature reduction with hydrogen iodide. 

The techniques of mineral dating by means of nuclear clocks have been 
described in previous Year Books. They are based particularly upon the 
measurement of rubidium-strontium and potassium-argon ratios. Research by 
L. T. Aldrich and G. W. Wetherill and their co-workers of the Department 
of Terrestrial Magnetism, and G. R. Tilton and G. L. Davis of the Geophysical 
Laboratory, have further extended and refined the results reported in previous 
years. Isotope analysis of samples from the Appalachian area of eastern North 
America and the southern Laurentian Shield in Canada was continued. Addi- 
tional age determinations were made for rocks from the Cutler batholith in 
Ontario, from upper Michigan and northeastern Wisconsin, from Oklahoma 
and Missouri, and from Death Valley, California. Analyses were also made of 
minerals from Finland, the European Alps, the Arabian Shield, western Aus- 
tralia, and the Venezuelan Andes. These mineral age determinations assist 
greatly in dating periods of orogenesis in the earth's crust. Pre-Paleozoic 
mountain systems in both eastern and western United States are gradually 
becoming delineated with their aid. 

During the year age determinations also were made of the ages of two or 
more coexisting phases in single rocks. Biotite, muscovite, and potassic feld- 
spars may be affected differently during orogeny and metamorphism. The 
research of Tilton and Davis on rocks in the Appalachian complex indicates 
that major geological events separated by very long intervening periods may 
be recorded by different minerals in the same rock. 

Further evidence on questions of mountain building has been obtained 
during the year in the course of the seismic program of the Department of 
Terrestrial Magnetism. In past years these studies have depended principally 
on seismic recording of mine and other man-made explosions. During the 
1957 reconnaissance in the Peruvian, Bolivian, and Chilean Andes it was 
observed that explosion waves showed unprecedentedly strong attenuation 
under both the Andes and the Altiplano of Peru. It is now planned to make 
use of local earthquakes for crustal studies. The first station of a local network 
for the detection of seismic disturbances planned for Peru, Bolivia, and Chile 
was installed near Arequipa during the year. These stations will be equipped 
with electronically operated seismic recorders requiring the minimum of atten- 
tion. It is further expected that evidence on subsurface structures extending as 
deep as 40 to 60 kilometers may be obtained from systematic and detailed 
observations resulting from large man-made explosions in the new open-pit 


copper mine at Toquepala, Peru. New information bearing on the depth of 
the Mohorovicic discontinuity and upon the orogeny of the Andes may be 

Four studies in other major areas of geophysical research during the year 
are of particular interest: the electric and thermal conductivity of the earth's 
crust, certain studies of radioactive fallout, research on the "electrojet" phe- 
nomenon, and observations of the radio emission of the sun. 

S. P. Clark, Jr., in the Geophysical Laboratory, has commenced a provocative 
study on the mechanisms of heat and electrical conduction of the earth and 
their relationships. Recent attention to the importance of heat transfer in the 
thermal regime of the earth has raised a new interest in its electrical conduc- 
tivity, since certain types of electrical conductors absorb strongly. The im- 
portance of inferences that can be drawn about the electrical conductivity in 
the earth from analysis of transient components of the magnetic field has been 
fully appreciated only recently, with advances in understanding of the mech- 
anism of electrical conductivity of the silicates, now known to be dependent 
on temperature. This incidentally raises the distinct possibility of developing 
a geothermometer based on electrical conductivity. Clark's studies are designed 
to improve our knowledge of the conduction process. His work during the 
past year was concentrated on the photon energy absorption of ferromagnesian 
silicates. He found that the strong ultraviolet absorption of iron-bearing 
olivenes was correlated with the presence of iron in the minerals, disappearing 
when iron was not present — an unexpected result, since previously the absorp- 
tion was thought to be intrinsic to the silicate structure. New and potentially 
important interpretations of the electronic structure of these silicates are being 
made from such absorption data. 

W. F. Libby served as a Research Associate of the Institution, working at 
the Geophysical Laboratory, during his entire term as Atomic Energy Com- 
missioner. During the past year he investigated the radioactive strontium con- 
tent of rainfall during the crucial period, March-May 1959, following an 
intensive series of bomb tests fired by the Soviet Union in October 1958. This 
event afforded a unique opportunity to test the characteristics of fallout from 
injections of nuclear materials at polar latitudes, as contrasted with the equa- 
torial explosions fired by the United States and the United Kingdom. Analysis 
of the strontium 89 -strontium 90 ratio in samples collected during the critical 
period showed that radioactive materials from the Russian polar shots had 
approximately a one-year time of residence in the stratosphere, as contrasted 
with the residence time of three years or more for radioactive materials injected 
equatorially. It is noteworthy that these analyses of the problems of Russian 
fallout, made in Libby's spare time from his Commission responsibilities, were 
actually the first to appear. Their value is evident both in connection with 


matters of public policy and as evidence for the patterns of the circulation of 
the upper atmosphere. 

The studies of the "electrojet" phenomenon in the high atmosphere at the 
magnetic equator, begun as a part of the United States program for the Inter- 
national Geophysical Year, were continued this year in the Department of 
Terrestrial Magnetism. This intense narrow band of electrical current varies 
sharply with local time, and especially during magnetic storms. The five 
observatories established last year in Peru as part of the Institution's IGY effort 
continued their observations during 1959, providing a full year of compre- 
hensive records for study. 

During the past two years several hundred scans of the radio emission of 
the sun have been made using the Department of Terrestrial Magnetism long 
antenna array. In addition, B. F. Burke and J. W. Firor have made some 
preliminary studies of all the nonthermal or "noise storm" events that have 
occurred during the past two years on the sun. A study of possible correlations 
with optical features of the sun was undertaken, but no correlation between 
the "noise storms" and optical features has so far been found. 

A notable event of the year in the field of solar studies in the Institution was 
a new departure in the program at Mount Wilson. Dr. Hale's lifelong major 
interest in the sun was vividly reflected in the Mount Wilson Observatory 
under his direction. The first instrument erected there for solar study was the 
Snow telescope, built in 1905. The 60-foot and the 150-foot tower telescopes, 
erected in 1907 and 1910 respectively, provided additional instruments, pio- 
neering for their time. Indeed, the Observatory was called the Mount Wilson 
Solar Observatory until 1918, when the 100-inch telescope was completed. 

A comprehensive study of the sun demanded a systematic set of observations 
scheduled for every clear day, because the observable features of its surface are 
of transient nature. The Observatories have maintained such a schedule for 
the past fifty years. In the course of it, about 23,000 direct photographs of the 
sun and more than a million spectroheliograms have been accumulated. They 
furnish a unique record of the numerous changes in the solar atmosphere that 
have occurred during more than four sunspot cycles. Much important infor- 
mation has been obtained about the motions, temperature, and composition of 
the solar atmosphere from study of these plates and related spectrograms. 
Early in the program these studies led to the discovery of the large magnetic 
fields of sunspots. Measurements of the strength and polarity of these fields 
were immediately included in the program of observation, providing Hale 
with the basis for his statement of the laws of sunspot polarities. 

Most of the projects for which these solar observations were originally under- 
taken have now been terminated. But the development of the solar magneto- 


graph by H. D. Babcock and H. W. Babcock has made possible the detailed 
observation of weaker magnetic fields which cover much of the solar surface. 
These widespread magnetic fields are variable, their strengths and distribution 
often changing rapidly from day to day. The investigation of these newly dis- 
covered fields provides a new frontier in solar research. To take advantage of 
it, the solar observational program was extensively revised on January 1, 1959. 
Many of the older types of observations were either discontinued or drastically 
curtailed, to free the time of instruments and of operating personnel for sys- 
tematic examination of these newly discovered magnetic fields. Another stage 
thus has been commenced in this long-continued program that has contributed 
so much to man's knowledge of the sun. 

One of the questions most persistently raised with astronomers, and of 
haunting interest to geologists, biologists, and the general public alike, is: 
does any form of life exist on planets of the solar system other than the 
earth? In general, visual inspection and study of direct photographs of the 
planets have not yielded as significant evidence on this question as indirect 
observations with thermocouples and spectrographs. Thus thermocouple meas- 
urements have shown that all the planets except Mars and perhaps Venus are 
at temperatures that would not permit of the existence of life as it is known on 

The planet Mars has received special attention in these queries because it 
appears to have surface temperatures and atmospheric conditions at least some- 
what similar to those of the earth. It has long been known that the surface 
of Mars exhibits large dark areas that move consistently with changes in 
season. There have been many speculations as to whether they represent 
shifting areas of plant coverage, or perhaps regions darkened by increased 
moisture content. Observations made some years ago by Dunham demon- 
strated that the amount of oxygen present in the Martian atmosphere is less 
than one per cent of our own, and the typical infrared reflection bands of 
chlorophyll have not been found in the spectrum of Mars. 

An important contribution to this disputed question about life on Mars was 
made during the report year by a guest investigator of the Observatories, 
W. N. Sinton, of the Lowell Observatory. Observations made by him in 1956 
with the 61-inch Harvard College Observatory telescope showed absorption 
bands of the integrated light of Mars in the far infrared at about 3.4 microns, 
which are ascribed to the methenyl group (CH). These bands appear in the 
reflection spectra of most plant life. To follow this promising lead further, 
Sinton was invited to continue his measurements with the 200-inch Palomar 
telescope. The large image obtainable at the coude focus of this instrument 
made it possible to examine the light and dark areas on the planet independ- 
ently for the presence of the infrared methenyl bands. Sinton's measurements 
definitely showed the bands to be much stronger in the spectrum of the dark 


regions than in that of the light regions. This is the most suggestive evidence 
yet to appear for the hypothesis that the dark areas on Mars are caused by 
some form of plant life which develops seasonally on these parts of its surface. 

The astronomy program of the year was also notable for the discovery of 
four new supernovae, three by Milton Humason and one by H. S. Gates, in a 
program directed by Zwicky. Two of those found by Humason were in near-by 
galaxies and were bright enough for detailed observation. Photometric obser- 
vations and spectrographic studies of them were started during the year. 

One of the major problems of present-day astronomy is the correlation and 
comparison of the properties of stars with differing masses, ages, and chemical 
compositions. Each of the globular and galactic clusters provides an assemblage 
of stars of about the same age and chemical composition, but with a wide 
range of masses. Different clusters, on the other hand, vary enormously in 
age (from perhaps a million to several billion years) and exhibit very large 
differences in chemical composition. Important information in these studies 
is the relationship between total radiation or magnitude and surface tempera- 
ture, as indicated by color index for a group of stars of the same age and 
composition. During the year investigations of color-magnitude relations have 
been made on stars within globular clusters, galactic clusters, and a galactic 
nucleus by H. L. Johnson of the Lowell Observatory, W. A. Hiltner of the 
Yerkes Observatory, W. A. Baum, G. Wallerstein, A. R. Sandage, and H. Arp. 

Evidence bearing on the chemical composition of stars has been sought in 
a project under the direction of J. Greenstein, supported by the Office of 
Scientific Research of the Air Force. The chemical compositions of a large 
number of stars have been studied, special attention being paid to differences 
in the abundances of elements between stars of different ages. All these obser- 
vations on color-magnitude relationships and chemical composition are being 
correlated with data on the generation of nuclear energy and on energy transfer 
from star core to star surface. It is beginning to appear from these studies 
that the magnitudes and other properties of stars of a given class are affected 
substantially by the large differences in chemical composition observed. This 
is an important — and a very disturbing — conclusion, since all distance, size, 
and mass determinations of distant objects such as clusters and galaxies have 
been based on the assumption of the uniformity in brightness of all stars of a 
given class, which, it now appears, may not be valid. 

The year also has been marked in the Institution by the development of 
new instrumentation for astrophysical observation. A notable event was the 
completion, as the year ended, of the new radiotelescope at the Derwood, 
Maryland, field station of the Department of Terrestrial Magnetism. This 
60-foot parabolic antenna reflector on an equatorial mounting was designed 
by the late H. E. Tatel. Though smaller than the 84-foot units installed at 
several locations in the United States and Europe during the past three years, 


it is an instrument of high precision which much increases the observational 
capacities of the radioastronomy group. It will be used particularly to extend 
the long series of measurements of the hydrogen clouds in our galaxy, made 
with the 54-channel H-line receiver and a 26-foot parabolic reflector. The 60- 
foot parabola will also be used in part to complement continuing work with 
the long antenna arrays, which during the past year increasingly demonstrated 
their capacity to yield precise positions of radio objects in the sky. Repeated 
measurements of known objects in Cygnus and Taurus have shown the errors 
of the long arrays to be less than one minute of arc. 

The year in astronomy was also marked at the Mount Wilson and Palomar 
Observatories by an event of another kind — the public presentation of the first 
successful attempts to photograph far-distant nebulae and galaxies in color. 
This spectacular accomplishment of W. C. Miller was made possible by his 
skill, by the unique sensitivity of Super Anscochrome color film, and by the 
exceptional optical speed of the 48-inch schmidt and the 200-inch Hale tele- 
scopes. Methods were perfected for the accurate determination and control of 
the shifts in color balance in the film emulsions during the long exposure times 
required for obtaining the pictures. The results, which were truly beautiful 
photographs, were displayed at the Carnegie Institution of Washington and 
the California Institute of Technology as large color transparencies. Repro- 
ductions also were published in the National Geographic Magazine and in Life. 

There has been further encouraging progress on what may well be one of 
the most important astronomical instrumentation developments of recent dec- 
ades. The Committee on Image Tubes for Telescopes, a joint project of the 
Institution, the National Bureau of Standards, and the United States Naval 
Observatory under the chairmanship of M. A. Tuve, Director of the Depart- 
ment of Terrestrial Magnetism, has reported some very promising tests for 
two types of image tube. A single-stage tube including a photocathode, pro- 
visions for appropriate electron optics, and a phosphor screen coated on mica 
was tested at the Lowell Observatory at Flagstaff, Arizona. Tests on the globu- 
lar star cluster M 3 showed a reduction in required exposure time by a factor 
of thirty as compared with direct photography using the same telescope and 
filter combination. A thirtyfold reduction in exposure time also was obtained 
in tests conducted on another type of image tube composed of two simple one- 
stage image tubes cascaded in series, with the phosphor screen of the second 
stage being photographed, using a fast relay lens. These results herald most 
important improvements of technique. 

Both types of tube still have shortcomings. The resolution of the single-stage 
tube is imperfect, and the cascaded tube produces a photograph disfigured by 
mottling. Both tubes include only a very small area in good focus. The devel- 
opment of still other types of tube is therefore being encouraged. The Com- 
mittee hopes to test the first samples of an improved cascaded image converter 


in late 1959. At the end of the report year arrangements also were being com- 
pleted to test two new image tubes with electrical output (television tube 
principle). Contracts are being maintained with five industrial laboratories for 
developing alternative types of tubes. The Committee is being greatly assisted 
by a grant from the National Science Foundation. 

The nature of life and of living processes is a major concern in the research 
program of the Institution. Throughout the world, research in these fasci- 
nating fields is expanding at a breath-taking pace. The connections between 
the worlds of geophysics, astronomy, and biology no longer seem as tenuous 
as they did only a few years ago. A wide range of amino acids, the building 
blocks of proteins, has been synthesized in the laboratory under electrical 
bombardment and ultraviolet illumination from materials, such as methane, 
ammonia, and hydrogen, thought to have been present in the primitive earth's 
atmosphere and hydrosphere, by Miller at the University of Chicago and, 
using a wider range of conditions and of raw materials, by Abelson in the 
Geophysical Laboratory. Outside the Institution, a protein has recently been 
produced synthetically from amino acids by Harada and Fox. Most spectacu- 
larly, what is clearly the key material of heredity, deoxyribonucleic acid, has 
been synthesized from the four "key" bases, the pyrimidines cytosine and 
thymine, and the purines adenine and guanine, plus a "starter," by Kornberg 
and his co-workers at Washington University at St. Louis. Also using an 
enzyme starter, Ochoa at New York University has synthesized RNA. It is 
reported that a synthetic DNA consisting entirely of thymine and adenine also 
has been made — a new kind of DNA, at least in the modern world of life — 
by Schachman. 

One of the most material — and the most vital — bridges between the radiant 
energy of space and the welfare of man is the still poorly understood phenom- 
enon of photosynthesis. All the higher forms of life on the earth today directly 
or indirectly depend on photosynthesis. This marvelous process has been a 
major subject of concern for the Department of Plant Biology for many years. 

Photosynthesis is dependent upon chlorophyll, of which two principal types 
have been recognized for more than a hundred years. ' Chlorophyll a is a pri- 
mary absorber for the conversion of light into chemical energy. The precise 
function of chlorophyll b is actually still unknown, after a century of specula- 
tion and experiment. During the year illuminating research has been addressed 
to this basic question at the Department. 

The late Robert Emerson, a former Research Associate of the Institution, 
recently showed that light absorbed by chlorophyll b strikingly increases the 
photosynthetic efficiency of long-wavelength light absorbed by chlorophyll a. 
In research undertaken this year at the Department, J. Myers and C. S. French 


have demonstrated that simultaneous absorption of light by chlorophyll a and 
chlorophyll b both increases the efficiency of light utilization by chlorophyll a 
and raises the initial photosynthetic rate after a dark period markedly over the 
situation when chlorophyll a alone is light activated. Enhancement effects 
are also observed from alternate pulsing, in periods of 0.6 to 15 seconds, of 
light sources of wavelengths absorbed by chlorophyll a on the one hand, and 
chlorophyll b on the other, suggesting that the enhancement results from 
interaction between the chemical products of two separate reactions initiated 
by light rather than from an interaction within the pigment system itself. 
Observations of the decline in rate of oxygen evolution following a period of 
illumination also indicate a difference in the products formed by chlorophyll b 
and by chlorophyll a. Although the specific chemical natures of the products 
formed by light acting on chlorophyll a have not been differentiated from 
those of chlorophyll b, it may now be concluded that chlorophyll b has a 
specific function in photosynthesis supplementing the better-known function 
of chlorophyll a, and does not act merely by transferring absorbed energy to 
chlorophyll a, as had earlier been surmised. 

The Department has also continued to study the characterization of the 
different forms of chlorophyll a occurring in living plants. It is now thought 
that there are at least three main types of naturally occurring chlorophyll a, 
distinguished by red absorption peaks at about 673, 683, and 694 millimicrons. 
The existence of these three different forms of chlorophyll a has been deter- 
mined principally by derivative absorption spectrophotometry of living algal 
suspensions. It is known that the proportion of the three types varies from one 
species of plant to another and also with conditions of culture. None of the 
components, however, has yet been separable by chemical extraction. It is 
thought that the three forms as they occur in the living plant result from 
different states of aggregation, adsorption, or chemical combination within 
the chloroplasts. 

During the year additional evidence was sought for the differentiation of 
the three chlorophyll a forms through fluorescence spectroscopy, differential 
bleaching, and in other ways. Of special interest was a series of measurements 
of the relative efficiencies of different wavelengths in photosynthesis in several 
species of algae, using equipment that gives an automatically plotted record 
of the action spectrum. The addition of constant supplementary light of spe- 
cific wavelength to the algae while the action spectra are measured provides 
a means for studying each of the pair of pigments whose simultaneous activa- 
tion leads to the enhancement effect of Emerson. 

Studies bearing on an understanding of the different forms of chlorophyll 
from the molecular standpoint were continued during the year by J. H. C. 
Smith and his co-workers in their investigations of protochlorophyll, its pre- 
cursor. Using high-speed centrifugation in a liquid medium of graded density, 


Smith and Coomber have obtained the protochlorophyll holochrome in a form 
relatively free from blue- and ultraviolet-absorbing contaminants with which 
it had always been earlier associated. These new preparations make possible 
the obtaining of improved spectra both for the protochlorophyll part of the 
holochrome complex in visible wavelengths and for the protein moiety in the 
ultraviolet. Protein determinations by ultraviolet absorption and by the biuret 
reaction now agree reasonably well, leading to a protein particle molecular 
weight of about one million associated with each protochlorophyll molecule. 

Smith and J. C. Goedheer have also studied the transformation from proto- 
chlorophyll to chlorophyll by observing the degree of depolarization in the 
fluorescence of freshly forming chlorophyll. Pigment fluorescence may be 
partly polarized when it is excited with plane-polarized light. As was reported 
in Year Book 57, a chlorophyll holochrome, formed by illumination of the 
isolated protochlorophyll holochrome, has a higher fluorescence polarization 
than chlorophyll in the mature living plant. The fluorescence undergoes de- 
polarization as chlorophyll accumulates during the greening of a leaf. Fluo- 
rescent light emitted by chlorophyll in mature living plants is only slightly 
polarized. Further experiments were undertaken during the year to investigate 
the reason for this differential. The harvested leaves of etiolated bean seed- 
lings, grown in the dark for nine days, were irradiated over periods up to 
seventy-five hours, samples being taken for analysis at a number of points dur- 
ing the time of illumination. Analysis showed a linear relation for the recip- 
rocals of fluorescence polarization plotted against chlorophyll absorbances. 
Values at the end of seventy-five hours of illumination agreed well with those 
obtained by other investigators on mature plants. These observations indicate 
that chlorophyll, as it accumulates, is attached to holochrome particles already 
containing chlorophyll rather than being joined to previously unoccupied 
carrier particles. Similar observations imply that chlorophyll molecules have a 
degree of rotational freedom within the holochrome particle, suggesting that 
the manner of inclusion of the chlorophyll molecule within the particle may be 
either through an amino acid "tail" such as tyrosine, histidine, tryptophan, 
proline, or oxyproline, or that the pigment is imbedded in an oil (fat)-water 
interface in which partial rotation is possible. There appear to be interesting 
clues here for an understanding of the different absorption spectra of chloro- 

In another interesting study, D. von Wettstein has explored the physical 
development of the interior of the chloroplast. His electron-microscopical 
studies of etiolated plastids have shown that chloroplast lamellae will form 
without the presence of protochlorophyll. This and a companion study suggest 
that chlorophyll is not a necessary component of the chloroplast lamellae 
proper. But the normal aggregation and close pairing of the lamellar disks re- 
quire the presence of chlorophyll. There is evidence that in the chlorophyll- 


containing grana regions of the chloroplast the pigment is spread in a mono- 
molecular film between the closely paired lamellar disks. Further electron- 
microscope studies have revealed differing submicroscopic organizations in dif- 
ferent cases. Chloroplasts have been observed in which the chlorophyll is not 
contained in lamellated grana regions as in the normal chloroplast. These and 
other clues suggest that the nature of combination of chlorophyll with different 
plastid components may be involved in the observed differences in the absorp- 
tion spectra of living chloroplasts. 

The year in the Department of Plant Biology also was marked by the isola- 
tion of a bacterial pigment having protochlorophyll-like absorption properties. 
R. Y. Stanier extracted the substance from a mutant of the purple bacterium 
Rhodopseudomonas spheroides. Analysis by Stanier and Smith showed the 
bacterial protochlorophyll to have an absorption spectrum similar to that of the 
protochlorophyll of squash inner seed coats, but with certain differences, such 
as the presence or absence of a phytyl ester group. Further study of this pig- 
ment should be of considerable interest in the general field of the biogenesis 
of photosynthetic pigments. 

At the very base of our understanding of the life processes — and indeed of 
the nature of life itself — lies a comprehension of the physical pattern in which 
the structures of living things are built, and the way in which the materials 
composing these structures are synthesized. Such inquiries are the chosen field 
of the Biophysics Section of the Department of Terrestrial Magnetism. 

A major concern of the Biophysics Section during the year continued to be 
the study of bacterial ribonucleoprotein particles. Similar particles, called ribo- 
somes, are found in all kinds of tissues, including those of bacteria, yeasts, 
molds, higher plants, and animals, suggesting their basic importance in life 
processes. It is known that they are concentrated where and when the rate of 
protein manufacture is highest, and it is believed that they provide the sites for 
most protein synthesis, including the formation of enzymes. It has been 
demonstrated that the capabilities of forming enzymes are dependent on the 
genetic constitution of the cell, and there is thus clearly an important relation 
between the ribosome and the chromosome in biosynthesis and the shaping of 
life structures and processes. 

Striking progress was made during the year in applying quantitative methods 
to ribosome problems. Of particular interest was the development of improved 
methods for physical separation of ribosomes of different sedimentation con- 
stants, by centrifuging through a sucrose density gradient, following a study of 
the conditions of stability of a liquid column in a gravitational (or centrifugal) 
field. By incorporating radioactive tracers in cells for very short periods, or 
"pulses," it has been possible to observe nascent protein which is associated 
with bacterial ribosomes for a short period before it is released as a finished 
product. Bacterial ribosomes are shown to have the same function in protein 


synthesis as the ribosomes of higher organisms, but the speed of synthesis by 
the bacterial ribosomes is about sixty times greater. 

The new measurements confirmed during the year the hypothesis that had 
been outlined in Year Book 56 as to the nature of ribosome particles and their 
growth cycle. Newly formed nucleic acid and protein appear in small particles, 
30S and 50S (sedimentation classes). These are the precursors of larger ribo- 
somes, 70S and 85S. There are about two thousand of each size per cell, one 
of each being synthesized on the average every two and a half seconds. Evi- 
dence has been obtained of a rapid circulation of materials among the various 
size classes, and of a cyclic relation between the large and small complexes. 
The 70S and 85S particles can be made to dissociate into the relatively stable 
30S and 50S ribosomes. The importance of concentrations of magnesium ions 
to the stability of the larger forms was clearly demonstrated, suggesting that 
ribosomes may be magnesium-dependent shaped objects. 

The nucleotide compositions of several classes of ribonucleic acid from 
Escherichia coli were determined by isotope dilution techniques, using chro- 
matographic and electrophoretic separations. The DNA of E. coli contains 
nearly equimolar proportions of its component nucleotides, and the bases are 
paired in a specific arrangement. Neither of these conditions appears to obtain 
for RNA. Thus there seems to be no simple correspondence between the com- 
position of DNA and that of any of the RNA's examined. On the other hand, 
there was shown to be a high consistency in the proportions of 6-amino-6-keto 
groups in the RNA types examined, so that the RNA composition seems to be 
under some restraint. 

Further information was obtained on the enzyme content of ribosomes. This 
included a determination that ribonuclease is a component of some, possibly of 
all, ribosomes; the identification of a new enzymic constituent of the ribosomes, 
leucine aminopeptidase; and measurement of the amount of beta-galactosidase 
associated with the ribosome. 

The mechanism of protein synthesis was also studied through the amino acid 
analogue technique, which may be said to induce "errors" in the synthetic 
process. In the analogue technique a foreign element may be substituted (like 
selenium for sulfur) or added (e.g., fluorine) to the amino acid molecule in 
the substrate in such a way that a growing culture of cells can incorporate it in 
place of the corresponding natural amino acid. The manner in which the 
analogue is incorporated has given considerable information on the mechanism 
of amino acid selection by cells. Experiments conducted during the year show 
that cells supplied externally with both natural amino acids and the correspond- 
ing analogues usually discriminate against the analogue as long as there is a 
supply of the natural acid. Once within the yeast cell, the ratio of analogue 
to amino acid remains about the same through the transfer of amino acids of 
the internal pool into the final proteins. Thus the process of selection appears 


to occur in the step by which amino acids enter the cell pool and less specificity 
is required in the "template" which arranges the amino acids into proper order 
in protein formation. Alternatively, the amino acids found in the internal pool 
may already have been selected by the template but not yet linked into poly- 
peptide strands. 

A closely related problem in biosynthesis continues to be attacked through 
another program of the Institution. The Department of Genetics also is 
heavily committed to the study of specific protein synthesis within cells and 
the mechanisms by which growth as well as replication or reproduction pro- 
ceeds. In the genetics program, however, the center of emphasis is the chromo- 
some and its nucleic acid content. Here are lodged the vital pilots for the 
origin, control, and continuation of life structures and life functions. Beyond 
the ribonucleic acid of the ribosome, the deoxyribonucleic acid of the chromo- 
some becomes a central point of interest. 

The studies of the Department of Genetics during the past year have in- 
cluded the chemical and structural composition of the chromosome, its replica- 
tion, and the transfer of the deoxyribonucleic acid content of the nucleus to 
the cell cytoplasm. Like the Biophysics Group, the Department has found the 
study of microorganisms a fruitful means of approach. Studies of both bac- 
teria and viruses were, as in several past years, prominent in the program of 
the Department; it also has continued the study of genetic problems by means 
of more complex organisms, principally the fruitfly (Drosophila) and maize. 

Further information on the molecular structure of chromosomes has been ob- 
tained by A. D. Hershey in experiments with bacteriophage, and by B. P. Kauf- 
mann and his associates in their studies of chromosomes treated with deoxy- 

Previous work by Hershey and G. Streisinger has shown that a particle of 
bacteriophage type T2 contains a single chromosome composed of nucleic acid. 
Analysis has indicated that there are about ten molecules of nucleic acid in one 
phage particle. Hershey has attempted to discover whether or not these ten 
molecules are identical. Experiments completed during the year indicate that 
they are identical both in glucose content and in chain length. If they are or- 
ganized in a tandem arrangement in the chromosome, as these findings suggest, 
the chromosome must be characterized by regularly spaced breakage points. 
Further investigation of this fundamental and fascinating point is now in 

Kaufmann and his associates continued their analysis of the responses of liv- 
ing cells to enzymes acting selectively on cellular components, with interesting 
results. When deoxyribonuclease enters the living cell a marked swelling and 
vacuolization of chromosomes occurs, which, however, can be inhibited by a 
prior treatment of the cells with ribonuclease. Thus much or all of the DNA 
in the chromosome might be degraded without loss of structural integrity. 


These observations suggest that DNA alone cannot be responsible for main- 
taining chromosomal integrity, which depends rather on the structural in- 
tegrity of the entire nucleoprotein complex. 

Other studies of the effects of deoxyribonuclease on Drosophila spermatozoa 
suggest that the enzyme can act as a potent mutagen. A new and discriminat- 
ing method of investigating chemical mutagenesis may now be in our hands. 

During the year Demerec and his colleagues succeeded in obtaining hybrids 
of Salmonella typhimurium and Escherichia coli, making it possible to com- 
pare homology of chromosome structure of these two bacterial species. Previ- 
ous experiments directed toward determining the sites of synthetic loci, using 
the phenomenon of transduction, have demonstrated that the genes controlling 
related biochemical reactions often are located in close proximity on the 
chromosome, in an order corresponding to the sequence of biochemical steps 
that they control. The transduction technique, however, allowed only small 
sections of the chromosome to be mapped at any one time. Study of the Sal- 
monella-Escherichia hybrids permitted the mapping of the entire bacterial 
genome. It has been shown that the genes controlling the synthesis of threo- 
nine, leucine, pantothenate, proline, and lactose have the same order of location 
in S. typhimurium and in E. coli. Three proline loci, and loci associated with 
the synthesis of glutamic acid and arginine, have been shown to lie close 
together on the Salmonella linkage map. These genes direct related biochemical 
reactions, and it must be supposed that their close proximity is biologically 
adaptive. Closely ordered arrangements of this kind have been observed only 
in bacteria. 

A very interesting series of experiments by Streisinger concerned the mecha- 
nisms of replication and recombination of genetic material in bacteriophages. 
Streisinger investigated the way in which the genome of bacteriophage type T4 
may interfere with replication of the bacteriophage type T2 genome in mixed 
infections. Mixed infections of single bacteria with phage T2 and a T4-T2 
phage multiple backcross have shown that the genome of T2 is partly elimi- 
nated from the progeny of these two. Experiments during the year were under- 
taken to discover whether the exclusion of parts of the T2 genome is a conse- 
quence of differing degrees of glucosylation in different sections of the genome. 
Analysis of phage particles with intermediate glucose contents suggested that 
the factors responsible for the degree of glucosylation were closely linked to the 
excluding factors, but the degree of glucosylation itself could not account for 
the observed exclusion. Further study is being made on the possible causes of 
this remarkable phenomenon. 

The functional relations between the chromosome and the cell whose struc- 
ture it controls, and of the chromosome's own internal regulating factors, con- 
tinue to be studied from very different viewpoints by H. Gay and B. McClin- 
tock. Gay's electron-microscope studies of the apparent physical transfer of 


chromosomal material, encased in nuclear-membrane blebs, from the nucleus 
to the cytoplasm in the salivary-gland cells of third-instar larvae of Drosophila 
melanogaster have been reported earlier. During the year she extended her 
observations to the salivary-gland cells of four other species of Drosophila — 
virilis, simulans, pseudoobscura, and funebris. The same phenomenon of 
transfer by blebbing was observed in all these species. 

McClintock has continued her cytogenetic studies of the suppressor-mutator 
system of control of gene expression in maize. The action of systems of genes 
controlling production of anthocyanin pigment in maize has been described in 
previous Year Books. Observations this year have disclosed some remarkable 
further features of the control. Spm elements differ widely in their activity. 
When an Spm in its active phase is incorporated into a nucleus carrying an 
inactive Spm that rarely reverts to its active phase, the inactive Spm is appar- 
ently activated. But it has been found that the inactive element suffers no 
permanent alteration in its association with the active one. Thus the integrity 
of an inactive suppressor-mutator element, even when an active element is 
incorporated in a nucleus with it, was established. The association of different 
Spm elements was found not to affect their subsequent control of phase re- 
versal; each retained its typical capacity. 

It is suggestive that there appears to be more than a superficial resemblance 
between the suppressor-mutator system of control of gene expression in maize 
and the system that controls the expression of flagellar antigens in the bac- 
terial species Salmonella, as described by Lederberg and lino. The modes of 
control, the changes in phase of activity, the duration of phase, and the "states" 
of the genes whose action is being controlled all bear a considerable resemblance 
in the two cases. It seems possible that we may have a suggestion of a remark- 
able parallel in gene control between a unicellular and a multicellular organism 
which are superficially very different. 

The biological work described thus far has largely centered in the structure 
and function of individual cells. Equally important, of course, is an under- 
standing of cellular integration and balance, of growth and development, in 
higher, multicellular organisms. This area of investigation, sometimes known 
as developmental biology, is particularly the concern of the Department of 

The discussion of the year's research in biophysics, photosynthesis, and ge- 
netics underlines the ever-increasingly intimate relation of biology, physics, and 
chemistry which characterizes modern investigation of living processes. James 
Ebert, the Director of the Department of Embryology, calls attention in his re- 
port to the growing conviction that embryology must be continually considered 
not only in the wider context of biology but also in that of chemistry and 
physics. If research explorations in biology are to converge, one natural point 
of focus is on the problems of development: how cell, tissue, and organism 


mature and are integrated. This terrain, familiar for the embryologist, now is 
attracting others who are concerned with molecular structure as a basic clue 
to understanding biological phenomena. The work of the Department of Em- 
bryology during the past year amply illustrates some consequences of this 

Studies of the mechanisms and mode of the acquisition of biological speci- 
ficity, such as those described last year, continue to be of central interest to 
the Department. This research is concerned with the synthesis and interaction 
of specific macromolecules, characterizing individual tissue types, individual or- 
ganisms within a species, and entire species. The Department has especially 
emphasized the study of tissue specificity. 

Gratifying progress has come through new experiments in immunoembry- 
ology employing nonliving tissue-specific antigens. It has previously been noted 
that most embryos and newborn animals are incapable of an immune response. 
A few days or weeks after birth, the antibody-forming system within an animal 
reaches a state of functional development, when antibodies are formed in small 
quantities. At birth, a population of mesenchymal cells exists which is capable 
of responding to an antigen. The ultimate differentiation of these cells results 
from a stimulation that can be controlled by the investigator. The capacity to 
recognize an antigenic stimulus and the capacity to respond to it have usually 
been considered inseparable in space and time. Ebert and L. E. DeLanney now 
suggest that there may be two stages in the development of the immune re- 
sponse, which may be experimentally separable. Before hatching, the chick 
embryo is incapable of a significant first-set homograft reaction. But these in- 
vestigators offer evidence suggesting that such a chick embryo can be sensitized 
by the antigenic stimulus of a first-set graft of adult heart. The embryo's spleen 
was shown to respond vigorously to a second stimulus as early as the fourteenth 
day. Ebert and DeLanney consider this response to be the stimulation only of 
the first stage of the immune reaction, when the antigen is "fixed," but before 
the antigenic message reaches the responding cell. Thus the role of the first-set 
graft may be the mobilizing of the initial receptors — an hypothesis readily 

These findings are puzzling in another respect. They appear to be contra- 
dictory to the prevailing concepts of actively acquired tolerance, with which 
embryologists and immunologists have been preoccupied for the past six years. 
Tolerance rather than sensitization might logically have been expected as the 
result of the experiments described. Two explanations can be advanced for 
the seeming contradiction: (1) immunologically competent cells were not in- 
cluded in the inoculum of the experiments, and (2) the grafts were implanted 
initially before the most favorable time for inducing tolerance. 

Among other significant questions arising from these experiments is the ef- 
fect of a mixture of antigens. When late embryos or newborn animals are 


exposed to a mixture of antigenic materials, what variations in response are to 
be expected? C. R. Wyttenbach has thrown additional light on this question 
by continuing his study of immune responses of newborn rabbits to injections 
of chicken liver and spleen homogenates. Normal rabbits were shown to pro- 
duce antibodies against these homogenates, but rabbits injected with the same 
type of homogenate at birth responded weakly or not at all upon subsequent 
challenge. Study of both the spleen-antispleen and liver-antiliver systems by 
the Ouchterlony serum-agar diffusion technique suggests that the animals not 
injected at birth respond strongly to a few major antigenic components, 
whereas the response of the treated experimental group toward these same com- 
ponents is eliminated or is reduced to the extent that the animals now respond 
instead to a variety of lesser components. 

Recent experiments by D. W. Bishop and R. Narbaitz support the expecta- 
tion that induction of tolerance in newborn guinea pigs differs from that in 
rabbits. Compared with rabbits, guinea pigs mature early. Probably the new- 
born guinea pig therefore has already passed through the optimal period for 
induction of tolerance. Bishop and Narbaitz have found that the injection of 
autologous testis and adjuvant into unilaterally castrated guinea pigs frequently 
depletes the semeniferous tubules completely of all spermatogenic elements. 
There is little, if any, recovery of spermatogenesis in these animals. The in- 
jection of the immature guinea pigs with mature homologous testis combined 
with adjuvant has also been found to induce aspermatogenesis. Animals less 
than ten days old injected once with mature homologous testis and adjuvant 
show complete or severe testicular damage when they are sacrificed as adults. 
This upholds the argument, contradicted by a previous negative report based 
on a small number of animals, that the injection of the newborn guinea 
pig with testis plus adjuvant should result in aspermatogenesis rather than 

Another interesting and significant investigation in the Department which 
was reported in a preliminary stage last year involves a study of the chemodif- 
ferentiation of visual pigments. F. H. Wilt has continued his work on the 
effects of thyroxine, the thyroid gland product which controls metabolism, 
upon the metamorphosis of visual pigment of the eyes of the bullfrog tadpole. 
The primary visual pigment of the larval eye of the bullfrog is porphyropsin, 
a vitamin A 2 -aldehyde-protein conjugate. During natural metamorphosis por- 
phyropsin is gradually replaced by rhodopsin, a similar conjugate of vitamin A x . 
The same change occurs in thyroxine-induced metamorphosis. Such artificial 
induction of metamorphosis in starved laboratory animals gave the first clue 
that thyroxine or its physiological derivative brought about this effect by in- 
ducing a fundamental change in vitamin A metabolism rather than by acting 
secondarily by causing dietary changes. Wilt's earlier studies on the organ 


specificity of thyroxine suggested, but did not prove, tiiat it acted directly on 
a rate-limiting factor in the eye itself. 

It is believed that a substantial part of this proof has been obtained during 
the past year. Thyroxine-cholesterol pellets were implanted either in one eye 
or in the abdominal cavity of the larva. The controls in the first case were con- 
tralateral sham-operated eyes of the same individual; in the second, they were 
nonoperated larvae. The relative proportions of vitamin A x and vitamin A 2 , 
an index of photopigment conversion, were determined for eyes directly ex- 
posed to thyroxine, eyes of animals carrying thyroxine implants in the ab- 
dominal cavities, control sham-operated eyes, and eyes from nonoperated con- 
trol animals. Improved methods of vitamin A isolation permitted a significant 
answer. The relative vitamin A x content was significantly greater in eyes di- 
rectly exposed to thyroxine than in those not so exposed. 

The phenomenon of embryonic induction, the specific influence of one type 
of tissue upon the embryonic growth of another, has been familiar for more 
than thirty-five years. Ribonucleic acid and ribonucleoprotein have generally 
been considered responsible for these inductive effects. Conclusive evidence, 
however, is lacking, despite a number of important recent studies like those of 
M. C. Niu of the Rockefeller Institute, and Toivonen and Yamada. Niu has 
reported that cultivation of ectodermal cells of an embryonic gastrula in ex- 
tracts of adult organs of the same species results in growth within the ectoderm 
of tissues embryologically resembling the source of the extract. For example, 
embryo cells of the amphibian gastrula treated with calf kidney extract have 
the appearance of tubule-like condensations, and calf thymus extract induced 
thymus-like structures in the ectoderm. Experiments by other investigators, 
however, have not always produced consistent results. Ebert has conjectured 
that many of the inconsistencies may result from a failure of the ribonucleic 
acid or ribonucleoprotein to reach its site of action. He has suggested that a 
combination of the presumed "inductor" with an RNA-containing virus might 
enable the inductor to reach the site more readily and so promote the effect. 
He has commenced a series of experiments designed to explore that possibility. 

The first experiments have dealt with the inductive ability of tissue micro- 
somes rather than of purified ribonucleic acid. The tissue to be induced was 
cardiac muscle, the site of inoculation was the chorioallantoic membrane, and 
the virus used was the RNA Rous sarcoma virus, which has an affinity for 
muscle, the source of the microsomes with which it was to be combined. Re- 
sults clearly indicate that inoculation of the chorioallantois of the chick embryo 
with a combination of Rous sarcoma virus and a cardiac muscle microsome 
fraction results in the growth within the membrane of cellular elements of 
both Rous sarcoma and muscle. No growths resulted from the inoculation of 
heart microsomes alone. Both Rous sarcoma virus and the combined Rous and 
heart microsome fraction, however, produced growths on the inoculated mem- 


branes. The two types of growth are indistinguishable grossly, but histological 
study reveals differences between them. Inoculation with Rous virus alone 
brings the characteristic ectodermal and mesodermal lesions often described. 
Inoculation with the Rous and heart fraction combination results in masses 
which contain in varying degree muscle or muscle-like elements intermingled 
among typical sarcoma cells. Although the incidence of clearly recognizable 
cross-striated muscle is low, large numbers of muscle-like cells and fibers are 
found among the tumor cells. They often resemble embryonic muscle, but 
occasionally they have the appearance of degenerating adult muscle. The en- 
couraging nature of these first results has given Dr. Ebert good reason for 
further exploration in this widely significant field. 

Studies having an ecological import, dealing basically with the interrelations 
of organisms with their environments, have long held a prominent place in 
the research program of the Institution. One of the important projects in this 
field, the assembly of comprehensive data for the study of range-grass hybrids 
of Poa by J. C. Clausen, W. M. Hiesey, and M. A. Nobs of the Department 
of Plant Biology, was finally completed during this year, and the analysis was 
started. With the cooperative support of the United States Department of Agri- 
culture, Clausen, Hiesey, and Nobs have perfected techniques of synthesizing 
new forms by interspecific hybridization in genera reproducing largely by 
apomixis. Ten new hybrid strains developed in the investigations appear, after 
extensive regional testing, to have potential agronomic value. The Poa program 
has contributed a most important set of basic data on the evolutionary mecha- 
nisms of a group of plants that has developed forms fitted to many different 
environments, and consistently stabilized them by apomixis. 

A similar study of genetic structure and evolution in the yarrows {Achillea 
sp.) by Nobs and Hiesey is nearing completion also. Current work continues 
to reveal evolutionary pathways that may have been followed in the develop- 
ment of the numerous existing forms of these plants. 

Most ecological plant studies of the past have been carried out under open 
field conditions. The newer pattern of such study is experimental growth un- 
der rigorous control of the temperature, humidity, and light of the environ- 
ment. In the numerous experiments undertaken in this country and abroad 
one important variable, however, has received little attention, namely the car- 
bon dioxide concentration of the air. During the year the Department of Plant 
Biology put into operation an experimental plant growth chamber, developed 
over several years past by Hiesey, Milner, and French, in which carbon dioxide 
concentration can be accurately regulated. This development gives to the 
Department a simple but highly effective facility for providing controlled 
conditions of temperature, humidity, light intensity, and carbon dioxide con- 


centration for precise ecological study. With the aid of this equipment, the 
comparative physiology of races of plants can be studied exactly, and reproduc- 
tion and replication of known plant environments can be provided where 
needed for quantitative measurements of photosynthesis and plant respiration. 
Hiesey and Milner are already making use of this approach in studies that 
quantitatively differentiate altitudinal races of Mimulus cardinalis. 

Another study of historicoecological interest was undertaken during the year 
by McClintock for the National Academy of Sciences in conjunction with the 
Rockefeller Foundation. It involved an analysis of the chromosome constitution 
of races of maize grown in Chile, Bolivia, Ecuador, and Venezuela as a cri- 
terion of their history in cultivation. Diagnosis was made of the presence or 
absence, size, and location of certain maize chromosomal characteristics, the 
"knobs." A relation was found between knob pattern and the geographic loca- 
tion of the race which appears specific enough to suggest that, under some con- 
ditions at least, a knowledge of the patterns of distribution based on chromo- 
somal-knob constitution may give valuable evidence of the origins and migra- 
tions and past history of hybridization of some early races of Indian corn. 

These studies suggest quite strikingly some of the directions in which eco- 
logical studies will find future opportunity. The precise measurement and 
description of the physiology of living organisms, so characteristic of today's 
biology, must be matched with equally precise studies in ecology if that field 
is to continue its meaningful development. Like other major fields, ecological 
study is also in transition, and that transition, like many others, is reflected 
within the Institution. 

Losses . . . 

Henning W. Prentis, Jr., member of the Institution's Board of Trustees and 
Chairman of the Board of the Armstrong Cork Company, died in Lancaster, 
Pennsylvania, on October 29, 1959. Faithfully attentive to the affairs of the 
Institution, Mr. Prentis served it as an active member of the Board from 1942 
until the day of his death. At the time of his passing he was a member of three 
Committees: Finance, Executive, and Terrestrial Sciences. He had served on 
the first for sixteen years, on the second for ten years, and on the last since 1957. 
He was also a member of the Committee on Biological Sciences between 1943 
and 1948, and a member of the Nominating Committee from 1947 to 1950. 
His devotion to the Institution was all the more remarkable because of his 
many other public-spirited affiliations and responsibilities. While directing the 
operations of a notably successful national and international corporation he 
found time to act as trustee or director for a dozen educational institutions, and 
he held more than twenty other important public service positions. A noted 
specialist on the techniques of management, Mr. Prentis always brought with 


him an interest in the future and a penetrating perception of human person- 
ality. His genial humor and kindly smile smoothed the course of many com- 
mittee meetings. They will be missed greatly. To paraphrase his own words, 
he held the secret of getting real joy and satisfaction out of life: the develop- 
ment of a passion for excellence for its own sake. 

Lindsay Bradford, vice-chairman of the Board of Trustees of the Institution 
from 1951 to 1953, and former president of the City Bank Farmers Trust 
Company of New York, died on October 6, 1959. The Institution benefited 
greatly from the perceptive mind of this distinguished financier. First elected 
a Trustee in 1940, he was a member of the Finance Committee every year 
between 1941 and 1958, and chairman of that Committee from 1947 to 1957. 
He was a member of the Executive Committee between 1954 and 1958. He was 
chairman of the Retirement Committee from 1954 to 1958 and had served as 
a member of the Nominating Committee in previous years. He resigned as a 
Trustee only last year. The outstanding success with which the Institution's 
portfolio was managed during his tenure on the Finance Committee owed 
much to his inspiration and guidance. Mr. Bradford was always an attentive 
observer of the Institution and its affairs, and his advice was much valued by 
its administrative officers. His contribution toward shaping the policy of the 
Institution during nearly two decades will be long remembered. 

For five years Dr. Willard F. Libby, in the midst of his arduous duties as 
an Atomic Energy Commissioner, served as a Research Associate of the Insti- 
tution. Few members of the Institution whose full energies could be given to 
their research exceeded his extraordinary productivity. At night, on Sundays, 
on holidays, Dr. Libby worked indefatigably at the Geophysical Laboratory. 
During those five years he wrote more than forty papers, at least twenty of 
which dealt with original research, much of which formed an essential back- 
bone — frequently the only essential backbone — of the great scientific questions 
in the consideration of which he played so major a role. New methods of 
hydrologic research were brought forward in tritium dating; studies of the 
absorption of beta rays, of the chemistry of labeled atoms, and of many facets 
of problems of fallout all were illuminated by his work. It will always remain 
as a striking demonstration of the fact that, by at least a few rarely gifted sci- 
entists, participation in great public issues, the administration of great enter- 
prises, and penetrating research work can be combined simultaneously. 

Dr. Libby has left Washington to return to research in the University of 
California at Los Angeles. It is well earned respite from the grueling years of 
public service, and it can only overjoy those who know him. But to the nation 
at large, to the Atomic Energy Commission, and to the Carnegie Institution of 
Washington that joy cannot but be tinged with a selfish regret for the loss of 
a quality of extraordinary rarity. 


In 1919 Samuel Callaway joined the staff of the Carnegie Institution of 
Washington in the Division of Publications. Within less than a year he became 
Secretary to the President of the Institution. In that capacity he served two 
Presidents: Dr. Merriam and Dr. Bush. During World War II he made an 
outstanding contribution to the nation as Secretary to Dr. Bush, first in Dr. 
Bush's capacity as Chairman of the National Defense Research Committee and 
subsequently as Director of the Office of Scientific Research and Develop- 
ment — services for which Mr. Callaway, in 1948, received the Presidential Cer- 
tificate of Merit. Then in 1953 he became Assistant to the President of the 
Institution, and continued to serve Dr. Bush, and later, until this year, myself. 

Mr. Callaway's services to the Institution spanned nearly forty years. He 
served three Presidents. He was, as I can eloquently testify, a rod and a staff. 
It is not possible to delineate the full meaning of his service, for it was too 
multivariant and complex, and too much an essential element of the Institu- 
tion. Likewise it is not easy to describe the gap that is left by his departure. 
But it is good to know that, although he may have retired from the Institution, 
he has by no means retired. On the contrary, he is at present actively in 
business in Virginia. 

Ten years after Mr. Callaway, in 1929, Dr. Alexander Pogo joined the Insti- 
tution as a staff member in the Division of Historical Research, to assist George 
Sarton in his work. For twenty-one years his knowledge of sixteenth-century 
science, Egyptian astronomy, Mayan chronology, and other fields, together 
with his extraordinary linguistic ability, served the Division. In 1950 he trans- 
ferred to the Mount Wilson Observatory, where he was Librarian and Editor 
until his retirement this year. As a result of his efforts the historical collections 
of the libraries at Mount Wilson and in Robinson Hall, Pasadena, have been 
strengthened. These two libraries now include one of the most complete collec- 
tions of early astronomical literature. Dr. Pogo's retirement from the Institu- 
tion, like Mr. Callaway's, does not mean his retirement. His remarkable 
linguistic ability will continue to serve others, as it has the Observatories in 
recent years. 

. . . and Gains 

It is a deep pleasure, once again, to report a number of honors which have 
come to members of the Institution. 

Three staff members or former staff members were elected during the year 
to the National Academy of Sciences: Dr. Philip H. Abelson, Director of the 
Geophysical Laboratory; Dr. Jens C. Clausen, retired staff member of the 
Department of Plant Biology; and Dr. Rudolph L. Minkowski, staff member 
of the Mount Wilson and Palomar Observatories. Three of the staff were like- 
wise elected to the American Academy of Arts and Sciences: Dr. Milislav 


Demerec, Director of the Department of Genetics; and Dr. Barbara McClin- 
tock and Dr. Alfred D. Hershey, both staff members of that Department. 

Dr. Hershey was also the recipient during the year of a joint Albert Lasker 
award of the American Public Health Association for 1959 for his leading part 
"in the discovery of the fundamental role of nucleic acid in the reproduction 
of viruses and in the transmission of inherited characteristics." Other par- 
ticipants in the award were Dr. Gerhard Schramm, of the Max Planck Institute 
at Tubingen, Germany, and Heinz Frankel-Conrat, of the University of Cali- 
fornia at Berkeley. 

Dr. Demerec was given an honorary membership in the Genetics Society of 
Japan "in appreciation of his outstanding contributions to the science of ge- 
netics." He has also been made an honorary member of the Sociedad de 
Biologia of Santiago, Chile, and of the Faculty of Medicine of the University 
of Chile. 

George W. Morey, retired staff member of the Geophysical Laboratory, 
whose contributions to the development of optical glass rank among the most 
famous of his many achievements, received the Howard N. Potts medal from 
the Franklin Institute of Philadelphia for his development of the glasses now 
used in lenses for high-speed photography. 

Ellis T. Bolton, staff member of the Department of Terrestrial Magnetism, 
received one of the Washington Academy of Sciences Outstanding Science 
Achievement awards jointly with Dr. H. George Mandel, pharmacologist of 
George Washington University. The award was made "in recognition of their 
outstanding studies on the biosynthesis of nucleic acids." 

Caryl P. Hasfyns 











Operated by Carnegie Institution of Washington 

and the California Institute of Technology 

Pasadena, California IRA S. BO WEN, Director 

HORACE W. BABCOCK, Assistant Director 


Ira S. Bowen, Chairman Jesse L. Greenstein 

Horace W. Babcock Rudolph Minkowski 

Robert F. Bacher Earnest C. Watson 



Introduction 43 

Observing Conditions 44 

Solar Observations 44 

Solar photography 44 

Sunspot activity 44 

Magnetic polarities 44 

Solar magnetic fields 45 

Stellar Spectroscopy 46 

Chemical composition of stellar atmos- 
pheres 46 

Subluminous and white dwarf stars . 48 

Novae 49 

Variable stars 50 

Magnetic stars 52 

Spectral energy distribution 53 

Stars with H and K emission lines ... 54 

Miscellaneous studies 55 

Gaseous Nebulae 55 

Globular and Galactic Clusters and Stel- 
lar Photometry 56 

Subdwarfs 58 

Absolute magnitude of RR Lyrae 

stars 59 

Absolute magnitude of cepheids 59 

Nucleus of a planetary nebula 60 

Galaxies 60 

The local group 60 

Rotation and dispersion of stellar 

velocities 61 

Stellar populations in ellipticals 62 

The distribution of luminosity 62 

Velocities of galaxies 62 

Supernovae 63 

Clusters of galaxies 63 

Instrumentation and New Observational 

Techniques 64 

Guest Investigators 66 

Staff and Organization 73 

Bibliography 74 

Carnegie Institution of Washington Year Book, 58, 1958-1959 


The Mount Wilson Observatory was 
planned originally for the study of the sun; 
indeed, until the completion of the 100-inch 
telescope in 1918, the official name was the 
Mount Wilson Solar Observatory. During 
the first decade of the Observatory's opera- 
tion, the Snow telescope and the 60-foot 
and the 150-foot tower telescopes were 
erected. These large fixed telescopes made 
it possible, for the first time, to fully exploit 
the capabilities of the spectroheliograph 
for the study of prominences, flocculi, sun- 
spots, flares, and other solar phenomena. 
The very powerful spectrographs attached 
to these instruments enabled Hale to dis- 
cover the magnetic fields of sunspots and to 
study their polarity and variations. 

Like meteorological phenomena on the 
earth, most of these solar phenomena are 
transient and change appreciably in time 
intervals from a few minutes to a few 
years. Systematic observations, regularly 
made, are therefore necessary for these 
solar studies. Almost from the start of 
operations on Mount Wilson a systematic 
set of observations was programmed for 
every clear day. This set included at least 
one direct photograph of the sun using a 
6-inch image, a spectroheliogram in Ha 
with the 6-inch image, spectroheliograms 
in both Ha and K with a 2-inch image, and 
another spectroheliogram exposed to show 
the prominences. In 1936, when the study 
of flares became important, an automatic 
spectroheliograph was installed. This took 
a spectroheliogram every 5 minutes 
throughout the day using a %-inch image. 
Later the rate was speeded up to give four 
spectroheliograms every 5 minutes. Daily 
tracings of the sunspots were also made 
from the 17-inch image at the 150-foot 
tower, and spot numbers were calculated 
on the basis of these drawings. The spec- 
trograph at this tower was used to make 
one or more determinations of the mag- 
netic polarity and field strength of each 

These photographs and drawings pro- 

vide a unique record of the sun for 50 
years, or more than four sunspot cycles. 
During this period, over 23,000 direct 
photographs and 17,000 spectroheliograms 
have been taken using the 6-inch image. 
At the same time, 105,000 spectrohelio- 
grams with the 2-inch image and 900,000 
with the %-inch image have been obtained. 
As the initial projects for the investigation 
of solar phenomena have been completed, 
and the original staff members of the 
solar department have retired, less and less 
use has been made of these routine observa- 

The development of the solar magneto- 
graph by H. D. and H. W. Babcock in the 
last few years has made possible the meas- 
urement of the strength and distribution of 
the weak magnetic fields that are usually 
present over much of the sun's surface. It 
has therefore become important to plot 
these fields at regular time intervals in 
order to study their fluctuations and their 
relationship to other solar phenomena. At 
the same time, Dr. Robert Leighton, of the 
Physics Department of the California In- 
stitute, has developed photographic tech- 
niques for the study of the detailed dis- 
tribution of the somewhat stronger mag- 
netic fields that accompany plage areas. 
He has also improved the techniques for 
the study of solar granulations. 

In order to free the time of the instru- 
ments and the operating personnel for the 
new studies of the magnetic fields of the 
sun, careful consideration was given this 
year to the drastic curtailment of the older 
types of observation. As the result of this 
study, the following changes have been 
made: (1) The routine flare patrol is dis- 
continued. (2) The large-scale direct photo- 
graphs and spectroheliograms and the spot 
drawings of the 17-inch image are made on 
alternate days only. (3) The classification 
of sunspots and determination of sunspot 
numbers have been placed on a low priority 
basis. These changes became effective on 
January 1, 1959, at the end of the Inter- 



national Geophysical Year and the com- 
pletion of the Observatories' responsibilities 
for supplying certain routine solar infor- 
mation for it. 

In order to increase the effectiveness of 
the instruments for the new programs, the 
Pyrex coelostat mirrors of both towers have 
been replaced by quartz mirrors. 


The precipitation on Mount Wilson for 
the year July 1, 1958, to June 30, 1959, was 
the second lowest on record, with a total 
of 17.63 inches compared with a 55-year 
average of 36.03 inches. Solar observations 

were made on 344 days, and observations 
were made on 315 nights with the 100-inch 
telescope and on 213 nights with the 60- 
inch telescope. 


Solar Photography in one calendar year. The previous record 

The daily program of solar observations ™ * 55 Z} 95 !' and th , e neX , t hi S hest was 

was carried out by Cragg, Hickox, and 663 ln 1947 ' Unprecedented sunspot ac- 

Seyfert. The numbers of photographs of ^ ^ du " n g * e re P ort L Y^ 

various kinds taken between July 1, 1958, * anuar y 1959 bein £ the hl S hest f °f 2 h ^ h 

j t ?n men t n a mean daily group number of 17.3. The 

and June 30, 1959, were as follows : ,., j • i i i <■ *r 

' highest during the last cycle was for May 

Direct photographs 648 1947 ? when the mean daily group number 

Ha and K2 spectrohehograms, 60- was 16#g< The evidence seems to in di ca te 

oo ocus that the maximum for the cycle was near 

Ha spectrohehograms, 18-root rocus. yoi , c £ , nM ,-p, J , , . 

K2 spectrohelio|rams 18-foot focus . 927 th f first L P art o£ 1958 / ™ e n ° rthern heml " 

K2 spectrohehograms, 7-foot focus. . 43,400 1 s P here has continued to be the more active 

K prominences, 18-foot focus 936 as > during the calendar year 1958, 493 

Magnetograms (regular disk and groups were observed in the northern 

polar scans) 167 hemisphere while 417 were observed in the 

southern hemisphere. 

Sunspot Activity More exceptionally high-latitude sunspot 

The magnetic classification and study of groups were observed in 1958, bringing to 

the sunspot and related phenomena were 33 the number of groups observed this 

continued by Nicholson and Cragg. As cycle at 40° or more from the equator, 

was mentioned earlier, these observations Only 12 such groups were observed in the 

were made on a reduced and somewhat ir- 80 years before 1954. 

regular schedule after January 1. Coopera- The monthly means of the number of 

tive programs have been carried out with groups observed per day for the past 2 x / 2 

the U. S. Naval Observatory, the Univer- years are shown in table 1. 
sity of Michigan, the Observatory of Kodai- 

kanal, the Meudon Observatory, the Cen- Magnetic Polarities 

tral Radio Propagation Laboratory, and the Magnetic polarities in each spot group 

Naval Research Laboratory. During the have, if possible, been measured at least 

calendar year 1958, solar observations were once. The classification of groups observed 

made on 339 days, on none of which was between July 1, 1958, and June 30, 1959, is 

the sun without spots. The total number indicated in table 2. "Regular" groups in 

of sunspot groups observed in 1958 was the northern hemisphere are those in which 

910, the largest number ever observed here the preceding members have N (north- 

1 These pictures in the form of a flare patrol seeking) polarity; in the southern hemi- 

were discontinued as of January 1, 1959. sphere the polarities are reversed. 



Daily Number of Sunspot Groups 





January 12.7 

February 10.9 

March 13.6 

April 14.0 

May 12.7 

June 15.1 

July 14.1 

August 12.4 

September 13.8 

October 18.9 

November 16.9 

December 19.7 f 

Yearly mean . . . 14.6 



















14.9 1 

* Based on 164 observed days out of possible 

t Largest monthly mean number of groups per 
day ever observed at Mount Wilson. 

t Largest yearly mean number of groups per 
day ever observed at Mount Wilson. 


Hemisphere Regular Irregular 


North 409 5 

South 251 4 

Whole sun. . 660 9 



Solar Magnetic Fields 

The fact that numerous stars exhibit 
magnetic fields, and the presumption that 
physically significant fields are ubiquitous 
among stars, emphasize the need for ade- 
quate physical models to account for the 
phenomena. The magnetic lines of force 
embedded in the fluid body of a star, and 
looping outward through its atmosphere, 
may be regarded as tracers by which the 
pattern of fluid circulation, oscillation, or 
turbulence can be resolved. This bears on 
the mixing of material at different layers 
within a star, on the ejection and accelera- 
tion of corpuscles above the surface, and 
on cosmogony. All these considerations 
revive interest in the problem of the 22- 
year magnetic cycle of the sun, which lies 

in the realm of magnetohydrodynamics. 
This problem is of peculiarly local interest, 
since it was at the Mount Wilson Observa- 
tory that George E. Hale discovered the 
magnetic nature of sunspots and the laws 
of variation of their polarity. It was Hale 
who instituted the search for the general 
magnetic field of the sun and who provided 
the stimulus and the equipment that have 
been fruitful for several decades in accumu- 
lating many of the basic data on the solar 
cycle. It is therefore appropriate, as noted 
in the introduction, that the orientation of 
the solar research program should be re- 
examined, and that it should be supported 
and amplified in those areas that pertain 
to magnetohydrodynamics. 

Two phases of solar investigation that 
are currently receiving attention are the 
study of the general or main dipolar mag- 
netic field of the sun by Harold D. Bab- 
cock, and the study by Robert Howard of 
the surface magnetic fields in active re- 
gions, with special emphasis on the rate of 
change of the details of the magnetic pat- 
tern and the relation of these local fields to 
solar flares and plages. 

In his continuing sequence of observa- 
tions of the general field, begun with the 
solar magnetograph in 1952, H. D. Bab- 
cock has discovered a striking phenome- 
non: the reversal of the main field. This 
reversal, which occurred first at the south 
pole and, more than a year later, in the 
fall of 1958, at the north pole of the sun, 
came near the maximum of magnetic ac- 
tivity in the current sunspot cycle. This 
reversal of the main field is probably a new 
aspect of the 22-year magnetic cycle, and it 
appears to be a finding of major impor- 

Howard observed solar magnetic fields 
with the magnetograph at the 150-foot solar 
tower. Using an aperture 10 seconds of 
arc on a side, he made a number of trac- 
ings of active and quiet regions. Contour 
maps, having considerably better resolution 
than the standard magnetograms, were 
plotted from the tracings. Several interest- 



ing discoveries about solar magnetic fields 
were made. Further evidence was found 
that within a few minutes of arc of a sun- 
spot there was generally enough magnetic 
flux of the opposite sign to balance that 
from the spot. In one instance, good evi- 
dence was found that the magnetic lines 
of force in the photosphere were tilted in 
the direction of a near-by spot by an angle 
that averaged 20°. The location and shape 
of magnetic "features" in the solar photo- 
sphere resemble very nearly those of cal- 

cium plages. The plages are outlined by 
about a 10-gauss contour line. In this study, 
filaments (prominences seen in absorption 
against the disk) always occurred between 
magnetic regions of opposite polarity. 
Some preliminary magnetic measures with 
an aperture approximately 1 second of arc 
in diameter indicate that the small-scale 
fields are not appreciably larger than those 
measured with larger aperture. Further 
observations with small apertures are 


During the past year 935 spectrograms 
were made with the 200-inch telescope, 822 
with the 100-inch, and 480 with the 60-inch. 

Chemical Composition of Stellar 

The project for the study of the chemical 
composition of stellar atmospheres has con- 
tinued throughout the year under the 
supervision of Greenstein with the finan- 
cial support of the Office of Scientific Re- 
search of the U. S. Air Force. 

The recent installation at Mount Wilson 
of a new grating providing a much higher 
dispersion (1.1 A/mm in the blue) has 
permitted exhaustive study of certain stand- 
ard stars. The analysis of these objects will 
require several years. Bonsack has taken 
spectra of c Virginis, G9 III, A4000 to A5000 
(1.1 A/mm) and A5000 to A6850 (1.7 A/ 
mm). He will obtain wavelengths, equiv- 
alent widths, curve of growth, and iden- 
tifications. A special goal is to provide ef- 
fective stellar line-absorption coefficients, 
which will supplement solar values of y\o, 
since the giant has stronger ionized lines, 
including rare earths. Bonsack intends to 
compare £ Vir with the weak-line, high- 
velocity G8 III star y Piscium. 

Saito has obtained similar spectra, X3400 
to A6900, for a Canis Minoris, F5 IV, and 
will make a detailed coarse and fine anal- 
ysis of its composition. Traving has ob- 
tained spectra of t Scorpii in the blue-violet 
region for a rediscussion of the profiles of 
the H and He I lines. Traving has also 

obtained spectra of y Serpentis, F5V, at 
4.5 and 7 A/mm to improve the data for 
a quantitative analysis now under way at 

Greenstein obtained at Palomar a series 
of spectra of the carbon-rich early R stars, 
HD 156074 (C 13 present), and HD 182040 
(C 13 absent, H and CH absent) for com- 
parison with the G8 III star 3 Herculis. 
An atlas of these spectra has been prepared; 
a detailed inspection and tracings by Bon- 
sack have been profitable. Comparison 
with laboratory spectra by Dr. R. B. King, 
of the California Institute, permitted elimi- 
nation of molecular lines. While C 12 C 12 is 
stronger in HD 182040 than in HD 156074, 
no C 12 C 13 is seen. The atomic lines in HD 
156074 are a good match to those in (3 Her. 
In HD 182040 many enhanced lines were 
found and identified with ions of normally 
abundant elements. Bonsack finds no spe- 
cial enhancement of the rare earths. In ad- 
dition, HD 182040 apparently has larger 
turbulence and may be more luminous 
than type III, although probably less so 
than R Coronae Borealis. The detailed in- 
vestigation of these stars will be carried 
out by Dr. Leonard Searle, of David Dun- 
lap Observatory, and Matsushima. Searle 
has completed and prepared for publica- 
tion an elaborate study of R CrB, based on 
Mount Wilson spectra. 

Greenstein has made a very careful spec- 
trophotometric study of the F2 V star o 
Bootis, as an absolute standard of equiva- 
lent widths. Plates from both Mount Wil- 


son and Palomar are available. The meas- 
urement procedure on tracings is point-by- 
point and unsmoothed; a total of 7 plates 
measured by Mildred Matthews gives ex- 
tremely accurate line profiles of Hy and 
H$ together with contributions of the 
weaker metallic lines. The stars G Leonis 
and 15 Vulpeculae have been added to this 

Heifer and Wallerstein are continuing 
to study the possible differences in abun- 
dances between young and old stars. In- 
cluded are HD 30455, A Aurigae, HD 
115043, C Herculis A, and 3 Comae Bere- 
nices. HD 115043 is a member of the Ursa 
Major cluster; HD 30455, X Aur, and C 
Her A are reported by Eggen to be mem- 
bers of high-velocity moving clusters. The 
comparison of C Her A and (3 Com is of 
particular interest because of the deviation 
of C Her A from the mass-luminosity law. 

The abundance of lithium in T Tauri 
and related stars has been investigated by 
Bonsack and Greenstein. Twelve stars 
were observed, mostly with 27 A/mm at 
the 200-inch coude. Strong Li I lines were 
found in T Tau, in RY Tauri (confirming 
results by Hunger), and in RW Aurigae, 
SU Aurigae, and GW Orionis. SU Aur is 
not classed as a T Tauri star by Herbig, 
but except for its higher luminosity it 
seems to be a member of the group. The 
following RW Aurigae variables did not 
show lithium: V 380 Orionis (a hotter 
star in a nebula), and CQ Tauri and PZ 
Monocerotis (not in nebulae). The pe- 
culiar high-latitude objects HD 117555, 
BD + 37°2318, BD + 24°2742, and AG 
Draconis have symbiotic or emission-line 
spectra and possibly blue continuous emis- 
sion. They are not in nebulae and do not 
show Li I. A rough curve-of-growth analy- 
sis gives an abundance ratio of lithium to 
the normal metals of about 100 times the 
value in the sun, equal to or higher than on 
the earth. The possible mechanisms for 
such an abundance increase were reviewed. 
Spallation by cosmic rays trapped during 
the early period of star formation is pos- 
sible; the possibility that lithium can be 

stored in the interstellar dust grains is also 

Work has been completed by Bonsack on 
a survey of the strength of the Li I reso- 
nance doublet (A6708) in cool normal stars. 
The abundance of lithium relative to va- 
nadium, or its upper limit, has been deter- 
mined for 46 normal stars of spectral classes 
G8 and M0, inclusive, and for the sun. The 
data for the stars were obtained from 
7 A/mm Mount Wilson coude spectro- 
grams, and those for the sun from the 
Utrecht Atlas and the observations of 
Greenstein and Richardson in 1951. No 
star was found with a lithium-to-vanadium 
abundance ratio significantly in excess of 
that in the sun. Stars of a given spectral 
class were found to have abundance ratios 
scattering from a certain maximum to val- 
ues as little as 1/100 as much; the maxi- 
mum value for each spectral class declined 
with surface temperature through a factor 
of about 100 from the hottest to the coolest 
stars studied. The abundance ratios at a 
given spectral class were found to be un- 
correlated with luminosity or any other ob- 
vious spectral characteristic. 

The apparent fluctuation of the strength 
of the Li I doublet relative to the other 
lines in the spectra studied indicates that 
the variation of the abundance ratio is due 
to changes in the lithium abundance. 
Greenstein and Richardson suggested that 
the deficiency of lithium on the sun rela- 
tive to the earth may be due to the circula- 
tion of the surface material to regions of 
high temperature by convective mixing. If 
this is true in cooler stars, the minimum 
efficiency of convection for the destruction 
of lithium increases with decreasing sur- 
face temperatures. Models computed for 
cool dwarf interiors have shown the pres- 
ence of convective zones at the surfaces of 
these stars, extending more deeply into the 
cooler stars. It is reasonable to suppose that 
a deeper convective zone destroys lithium 
more effectively. The lithium abundance 
results then suggest that such convection 
zones also exist in stars of higher luminos- 
ity, and that the depth of convection is 



cither not the only parameter affecting the 
efficiency of lithium destruction or that the 
depth varies strongly from star to star, 
above a certain minimum value, in stars of 
given surface temperature. 

Traving has analyzed spectra of an in- 
teresting faint, high-latitude B star, BD + 
33°2642, obtained by Greenstein. This star 
is at a very great height above the galactic 
plane and was suspected of being a halo- 
population B star, like Barnard 29 in the 
globular cluster Messier 13. The spectrum 
resembles that of t Sco, except that the 
Balmer series is sharper in the faint star. 
The relative concentration of H in the sec- 
ond level (, and of He I (N 2 3ph) 
was determined by (1) assuming an op- 
tically thin layer, and (2) using saturated 
lines broadened by ions (statistical theory) 
and electrons (collisional theory). The 
electron density log N e = 13.32 and log 
N ,2h/N 2 Sph = + 1.06. The stratification in 
the atmosphere was neglected, but allow- 
ance was made for the change of the effec- 
tive thickness of the atmosphere, h. The 
abundances of H, He, and other elements 
are being determined. The same methods 
were applied to t Sco. The results are 
log 2V„= 14.44, log No.2h/N 2 3ph=+0.%. 
The halo star thus has a lower electron 
pressure and about the same helium abun- 
dance as t Sco; it may be above the main 
sequence (or have a low surface gravity). 
The abundance of the heavier elements 
seems to be low. 

The effect of the source of opacity on 
the change in equivalent widths of metal- 
lic lines, as a function of metal content, 
has been discussed by Saito. He studied 
two sequences of models with different 
H/metals ratio, i.e., log ^4 = 3.8 and 5.8. In 
late F and G subdwarfs the electrons come 
from hydrogen; in late K and early M 
stars the metals supply the electrons and 
so the opacity depends on the metal abun- 
dance. If H~ is the opacity source, Y] is 
reduced and lines are weakened by a de- 
crease of metal abundance in F and G 
subdwarf stars, but not in K and M stars. 
In late-type metal-poor stars, the increase 

in pressure broadening will actually en- 
hance strong lines. Rayleigh scattering 
may have a role in extremely metal-poor 
stars. (Traving suggested this for the red 
giants of Population II.) Molecules, de- 
pending on the H/O and H/metals ratio, 
may be different in strength in late-type 

Subluminous and White Dwarf Stars 

Many white dwarfs have been found 
or confirmed spectroscopically by Green- 
stein from the lists of faint blue stars. They 
include Feige 24, 26, 34, 43, 93, HZ 2, 4, 
14, 28, L 710-30, L 845-70, LB 1240, and 
Tonantzintla 547. 

A number of planetary nebulae of very 
low surface brightness were found by 
Abell. Greenstein has obtained spectra of 
seven, and Minkowski of one, of the stel- 
lar nuclei. The nebulae are so faint that 
there is no interference from emission lines. 
Using Shklovski's method, Abell has es- 
timated luminosities of the central stars 
from Mpg—0 to 4-8. These low luminosi- 
ties are confirmed by the spectra. Lines 
of H, He II are very broad and shallow, 
and the stars of the lowest luminosity re- 
semble hot white dwarfs, although with 
weaker and somewhat narrower lines. 
These stars fill the gap from the horizontal 
branch to the white dwarfs. One object, 
Abell 78, is noteworthy in having rela- 
tively high luminosity and Of-type lines. 
It has, in addition, a strong double emis- 
sion at A3810 and A3835, which is not pro- 
duced in the large nebula, and arises from 
O VI. 

Various types of hot subluminous stars 
in the halo population have been studied. 
A number of very hot sdO stars have been 
found, in which He II and He I are ac- 
companied by weak, broad H lines. In 
some very blue stars with strong H lines, 
a weak diffuse A4686 is often found. 
Among the sdB stars, H is very strong and 
the He I lines vanishingly weak. The cor- 
relation between type and color is not 
perfect, and it appears that some hot blue 
stars show nothing but H lines. Ordinary 



halo A and B stars are found very com- 
monly among the blue stars in lists by 
Feige, Haro, and Luyten. A number, but 
not all, of these stars have high velocity. 
Several composite blue stars showing H, 
He I, and Ca II have been found; these 
are possibly sdG and sdA pairs. 

A preliminary survey of K and M sub- 
dwarfs has been made to study the effects 
of possible low metal abundance on these 
spectra. The most striking phenomenon 
is the great strength of the MgH bands in 
the sdK6 star Ross 451. In the late M stars 
differences are quite small. 

Traving has applied the theory of reso- 
nance scattering of hydrogen lines to the 
sun, and of helium lines to the white 
dwarf L 1573-31. He has suggested a very 
interesting charge-transfer collisional proc- 
ess as a possible explanation of the uniden- 
tified band at A4670 in the white dwarf 
W 219, i.e., H 0=2) 4- H = l)-»H-4- 
H + . 

Saito has computed model atmospheres 
for white dwarfs of types DA, DB, and 
DC. The absorption coefficients were ob- 
tained for the range of temperature 6 = 0.2 
to 0.5, log P e 0.0 to 6.0. Pure helium model 
atmospheres with surface gravity log g—S 
and 9, e = 0.252, 0.335, 0.470, were con- 
structed. In addition, six models with the 
normal H/He ratio were computed. The 
U — V and B — V magnitudes, including 
blanketing effects, were computed. The 
(B — V, e ) relation obtained is not far dif- 
ferent from Greenstein's approximate val- 
ues for the white dwarfs. For pure helium 
stars, e =O.252, 0.335, 0.470 give B-V 
colors of —0.20, —0.15, and 0.00 magni- 
tude. The standard mixture gives B — V 
colors —0.20, —0.05, and 4-0.15 magni- 
tude for the same temperatures. 

The broadening of Hy and He I singlet 
and triplet lines in white dwarfs has been 
studied by Saito. The Holtsmark theory 
gives agreement with the observed depend- 
ence of the equivalent width of Hy on 
temperature. The slow decrease of W 
(Hy) for 0<O.3 is caused by the decrease 
in the effective population of the second 

level. The sharp decrease for 0>O.55 is 
caused by lowered ionization and the re- 
duced N e . (In main sequence stars N e 
drops more slowly with decreasing tem- 


The outburst of the recurrent nova RS 
Ophiuchi on July 13, 1958, was observed 
by Wallerstein on the first night and on 
12 of the succeeding 16 nights by Deutsch, 
Greenstein, Oke, or Wallerstein. Observa- 
tions were made by Joy on 12 nights be- 
tween August 5 and October 20. On the 
early spectra, some of which were taken 
at dispersions as high as 9 A/mm, many 
sharp emission and absorption lines with 
only a small velocity shift were superim- 
posed on the broad emission features. Wal- 
lerstein has interpreted these as indicating 
that RS Oph possessed a circumstellar 
envelope before the outburst. 

With the lower dispersions used by Joy 
to observe the later phases, no absorption 
lines were seen, but emission lines, usually 
4 to 5 A in width, appeared in increasing 
numbers during the period of observation. 
In addition to the emission lines of H and 
He, which are 10 to 20 A wide, more than 
200 lines were tentatively identified by 
measures of wavelength as possible emis- 
sion contributors. 

Iron lines are the most numerous, Fe II 
and [Fe II] being nearly equal in number. 
Fe II lines are generally stronger and 
reach their maximum intensity earlier. 
Forbidden lines of highly ionized iron 
atoms from [Fe V] to [Fe XIV], whose 
intensities increase with time, also occur, 
the strongest being A 6374 [Fe X]. Other 
highly ionized atoms represented with rea- 
sonable certainty by forbidden lines are 
Ne, S, A, K, Ca, and Ni. The strong line 
A6827 previously attributed to [Kr III] was 
first seen at phase 31 days, and at the last 
observation it was comparable in strength 
with the strongest He I lines. No more 
likely identification has yet been suggested, 
although the absence of the other forbid- 
den Kr III line at A5423 is disturbing. 


Permitted lines of C III, N II, N III, and 
Si II have their greatest intensity at phases 
near 35 days, but no Ti lines are present. 

In general, the results obtained in 1958 
lead to the conclusion that in the recent 
outburst of RS Oph the spectroscopic be- 
havior of 1933 was repeated with great 
precision. The detection of many more 
faint lines in 1958 is probably the result of 
improved spectroscopic equipment. 

An orbit for the nova DQ Herculis has 
been obtained by Dr. Robert Kraft, of 
Yerkes Observatory, and Greenstein on 
the basis of spectra obtained by Greenstein 
at the prime focus of the 200-inch tele- 
scope. Observations were made around the 
cycle and, in addition, very short exposures 
during the eclipses. The Balmer decrement 
varies with phase because of different con- 
tributions to higher members of the series 
by the star and its surrounding disk in 
comparison with the envelope. The He II 
line, A4686, is present mainly in a nonuni- 
form disk surrounding the nova, and dur- 
ing eclipse varies in velocity by about 900 
km/sec in only 2 minutes. A large rota- 
tional effect of about 300 km/sec is super- 
posed on the orbit determined from A4686; 
the orbital semiamplitude is about 150 km/ 
sec. The hydrogen lines show a complex 
structure. The mass of the nova, depend- 
ing on the mass ratio, is about 0.3 to 1.0 
solar mass; the lower value is consistent 
with the nearly white-dwarf nature of the 

Variable Stars 

The spectrum of the pure S-type long- 
period variable star R Cygni (P=430 days) 
has been studied at two successive maxima 
by Deutsch and Merrill. At the abnor- 
mally low maximum of 1957, the spectrum 
exhibited an unprecedented array of hun- 
dreds of sharp metallic emission lines. 
These represented all the expected transi- 
tions in abundant atoms and ions, with 
the conspicuous exception of those to the 
ground level, virtually all of which were 
missing. At the brighter-than-average 
maximum of 1958, however, the usual S- 

type absorption-line spectrum was shown. 
From a discussion of these spectroscopic 
peculiarities and the associated radial-ve- 
locity differences, Deutsch and Merrill 
have been led to a complex, stratified 
model of the R Cyg atmosphere. In par- 
ticular, it appears necessary to conclude 
that the temperature not only rises out- 
ward from the photosphere, as in the 
solar chromosphere, but that it then 
reaches a maximum and falls again to 
very nearly zero in the circumstellar en- 
velope. It may be possible to understand 
some of this structure in terms of the 
effects of a succession of shocks, which 
both heat and levitate the atmosphere. 
Deutsch and Hans Liepmann, of the 
California Institute, are attempting a theo- 
retical discussion of some of these phe- 

Merrill has completed an investigation 
of the behavior of the spectrum of AG 
Pegasi during the years 1949 to 1956. 
Within the past 40 years this object has 
developed into a fairly typical symbiotic 
star. The recent behavior of the forbid- 
den lines of O III is of special interest. 
The observed motions and widths of these 
lines may arise from a jet of gas of rela- 
tively high density emitting chiefly A4363, 
revolving near the center of the system 
and discharging into an expanding outer 
zone which emits chiefly A5007 (see Astro- 
phys. J., 129, 44, 1959). 

From spectrograms with dispersions 
from 2.3 to 4.5 A/mm of the long-period 
variable o Ceti, type M6e, Merrill has com- 
piled a list of 1347 absorption lines, mostly 
atomic, between A3528 and A4462. Meas- 
ured displacements increase with the 
atomic excitation potentials of the lower 
levels of the lines by 1.3 km/sec per elec- 
tron volt. The probable cause is an upper 
expanding shell of gas cooler than that in 
the main reversing layer. 

V725 Sagitarii, which changed bright- 
ness slowly from 1889 to 1928, began to 
vary in light with a period of 16 days in 
1928, the period, according to Miss Swope, 
increasing at the rate of 1 day per year 


until 1935. No observations of the star 
have been published since that time. Two 
spectrograms (170 A/mm) obtained by 
Preston with the 100-inch show G-type H 
and K lines, weak hydrogen lines, and ex- 
tremely weak lines of other metals, the 
spectrum resembling in these respects a 
red giant in M 15 observed with the same 
equipment. The star still shows indications 
of variability and is being followed photo- 
metrically by Dr. L. Plaut, of the Kapteyn 
Laboratory at Groningen, and Miss 

The metallic-line spectrum of the irregu- 
lar variable HH Monocerotis yields a type 
of approximately F6 IV. The H and K 
lines of Ca II, however, are those of an 
A7 star. A spectrogram well exposed in 
the ultraviolet shows no other indication 
of an A-type companion. The object may 
be related to the metallic-line stars. A 
search by Preston for other stars of this 
kind among the irregular and suspected 
variables known to have F-type spectra 
has, so far, proved futile. Of the eight 
stars observed, six appear to be normal 
main-sequence stars, one is an F super- 
giant, and one, V425 Cygni, is a B-type 
shell star with diffuse helium lines, sharp 
Balmer lines that can be counted to H24 
at 80 A/mm, and a strong narrow emis- 
sion at H3. 

Spectroscopic observations (80 A/mm) 
of RV Ursae Majoris, an RR Lyrae star, of 
Bailey type a, have been obtained by Pres- 
ton with the 60-inch telescope from Febru- 
ary through July 1959. The light-ampli- 
tude of RV UMa is strongly variable in a 
period of 90 days. Radial-velocity measure- 
ments of about half of the material indi- 
cate that the velocity amplitude is also 
variable, ranging from 40 to 70 km/sec. 
The phase of velocity minimum in the 
fundamental period advances when the 
velocity amplitude is large, as in the case 
of XZ Cygni. The star has been observed 
photoelectrically in Berkeley and at Lick 
Observatory by Mr. Hyron Spinrad, and 
the spectroscopic and photometric data will 
be discussed jointly. 

Less extensive coverage of the variables 
of Bailey type a, SU Draconis, SW Dra- 
conis, AR Persei, TU Ursae Majoris, UU 
Virginis, as well as scattered observations 
of several more variables, has been ob- 
tained. This set of stars possesses the 
entire range in K-line types previously 
found at very low dispersion among the 
field variables. Weaker metal lines that 
can be seen at 80 A/mm share the anom- 
alous weakening of the K line, but to a 
lesser extent. This can be understood 
qualitatively as a curve-of-growth effect. 
The weaker features are composed for 
the most part of lines that lie on the flat 
part of the curve of growth and are con- 
sequently less affected by abundance varia- 
tions than the K line, which has strong 
damping wings in all the cases considered 
here. It is doubtful that an unambiguous 
separation of temperature, surface gravity, 
and abundance effects is possible from a 
study of 80 A/mm spectra of these stars. 

Spectrograms with a dispersion of 20 
A/mm are being collected for the purpose 
of making a comparative study of a small 
number of RR Lyrae stars with strong, 
intermediate, and weak lines. A pre- 
liminary line identification is now in prog- 

Spectrograms with a dispersion of 400 
A/mm of RR Lyrae stars in M 5 and 
M 92 (Oosterhoff types I and II, respec- 
tively) have been obtained with the 100- 
inch by Preston. The M 92 variables have 
AS values (hydrogen-type minus K-line 
type) of approximately 10; those for the 
M 5 variables are much smaller. Thus, the 
spectroscopic differences detected by 
Deutsch in red-giant spectra from cluster 
to cluster can be found among horizontal 
branch objects as well. A mixture of vari- 
ables from these two clusters appears to 
be capable of providing the spread in spec- 
troscopic properties observed among the 
weak-line RR Lyrae field stars. An earlier 
failure to detect such differences between 
M 5 and the Oosterhoff type II cluster 
M 22 may be due, in part, to the inter- 
mediate line-weakening of M 22 indicated 



by Deutsch's spectrograms of giants in 
that cluster. 

The 4-inch camera of the X spectrograph 
of the 60-inch telescope was used by Var- 
savsky to determine the radial velocities 
of three RR Lyrae-type stars, namely, SU 
Draconis, VY Serpentis, and AP Serpentis. 
Both observations and reductions are com- 
plete. The 60-inch telescope was used to 
determine three-color light-curves of three 
RR Lyrae-type stars, namely, TU Ursae 
Majoris, AP Serpentis, and VY Serpentis. 
The observations have been completed; 
part of the reductions have been done by 
hand, and the rest are under way using 
Arp's program for the Datatron machine 
at the California Institute. 

Magnetic Stars 

Magnetic fields of numerous stars are 
under investigation by H. W. Babcock, 
mostly with the 200-inch Hale telescope, 
with the dual aim of refining the data for 
a few of the outstanding variables of large 
magnetic amplitude and of broadening the 
basis for statistical discussion by observing 
a greater number of somewhat fainter stars 
with reduced precision. For the second 
phase of the work, the 36-inch camera of 
the coude spectrograph permits reasonable 
exposure times on stars as faint as the 
ninth magnitude. 

Some of the fainter objects now being 
investigated are identified by the follow- 
ing numbers in the Henry Draper Cata- 
logue: 18078, 115078, 192678, 135297, 171586, 
171782, 171914, 191742, 200311, 215038, and 
215441. This program is designed to pro- 
vide answers to such questions as the fol- 
lowing: What are the relative numbers of 
a, 3. and y variables? Do all a variables 
have large amplitudes, and what is the 
distribution function of their periods ? Do 
all periodic magnetic variables show re- 
versal of polarity? Is the rate of magnetic 
variation connected with the depth of the 
hydrogen convective layer or with other 
physical parameters of the star? 

Among the brighter magnetic variables, 
it has been found that HD 32633 is irregu- 

lar; the star shows a very strong field with 
rapid fluctuations, reversal of polarity, and 
a uniquely large crossover effect. The AOp 
star HD 187474, which for a least a year 
was suspected of having a constant field 
of about —1800 gauss, has been found to 
be in fact a very slow variable. The latest 
observation shows that the field has de- 
creased to about zero. Thus, no star is 
known to show a constant magnetic field. 

A unique transitory effect was observed 
on one spectrogram of HD 170901, which 
showed some 60 weak sharp lines of the 
metals in left-hand but not in right-hand 
circularly polarized light. These lines ap- 
peared on only one plate; numerous other 
spectrograms show only an ordinary A- 
type spectrum with broad hydrogen lines 
and a few rotationally broadened shallow 
lines such as A4481. This peculiar effect 
can be attributed to temporary promi- 
nence-like activity, in which, under very 
special circumstances, the atoms were 
aligned in a magnetic field and thrown 
into preferred magnetic sublevels by ab- 
sorption of microwave resonance radiation. 

In the AOp star HD 125248, H. W. Bab- 
cock has shown that the effective magnetic 
field varies between —1900 and +2100 
gauss with a period of about 9.3 days. Syn- 
chronous variations occur in line strength 
and radial velocity. Deutsch has studied 
the variations of the lines of Eu II, Gd II, 
and Ce II which show one characteristic 
pattern of variation; those of Cr I and 
Cr II which show a second, different pat- 
tern; and those of Fe I, Fe II, and Ti II 
with a third pattern. He has continued 
his efforts to represent all the observed 
changes by means of a rigid-rotator type of 
model, in which the geometry of the spec- 
troscopic patches and of the associated 
magnetic field is derived from the observed 
variations of He, radial velocity, and line 

One obstacle to a satisfactory discussion 
of this kind is the presence of a long- 
period velocity variation in the star, pre- 
sumed due to Kepler motion, of ampli- 
tude large compared with the variation in 


the 9.3-day cycle. To eliminate this sec- 
ondary velocity variation from the analysis 
of the velocity curves, Deutsch has now 
discussed together all the 55 coude spectro- 
grams of HD 125248 that were obtained in 
the 11 years before 1957.2 and have been 
measured for radial velocity. 

The velocities obtained from each of the 
three groups of lines have been analyzed 
for the (common) orbital elements of the 
spectroscopic pair, and for the five addi- 
tional parameters needed to describe the 
group's 9.3-day velocity curve. The method 
of least squares was applied, by means 
of the Datatron 205 digital computer at 
the California Institute, to the suitably 
weighted observed group velocities, which 
numbered 129. After several successive ap- 
proximations, the weighted sum of squares 
of residuals on this model was reduced by 
more than 5 times, and the representation 
of the observations could not be signifi- 
cantly improved. The external probable 
error of a velocity derived from about 25 
absorption lines in one group was found 
to be only 0.7 km/sec. The extensive com- 
puting required in preparing the material 
for the Datatron was done by Miss Burd, 
and Miss Lowen undertook all the work 
at the Datatron itself. 

From this discussion, Deutsch has drawn 
the following conclusions, among others: 
(1) After correction for Kepler motion 
with a period of 1618.0 ±8.1 days, and an 
amplitude K of 7.60 ±0.17 km/sec, the ob- 
served velocity curves in the 9.3-day cycle 
remain stable over the 11-year period cov- 
ered by the observations; and within the 
precision of measurement they repeat 
themselves from cycle to cycle. (2) There 
is little or no evidence for discontinuities 
in these velocity curves, or for sudden ac- 
celerations. (3) The phase relations be- 
tween velocity and intensity variations 
within each group are precisely those re- 
quired for a rigid-rotator model. (4) The 
amplitudes of these curves are compatible 
on this model with the widths of rela- 
tively nonvariable lines, provided that the 

star is viewed at an inclination of about 
30° to the rotation axis. 

Spectrograms are also being obtained by 
Deutsch for a similar mapping of Bab- 
cock's A2p "Alpha variables," 53 Camelo- 
pardalis and HD 98088. In the former 
star, intensity estimates show that the lines 
of Cr I change strength in antiphase with 
those of Cr II. The Cr II lines are strong- 
est and widest near Ti II maximum, about 
6.5 days after positive crossover, along with 
lines of Fe I and II, Sr II, and Si II, and 
they are sharpest and weakest half a cycle 
later. The intensity of Mg II 4481 shows 
a double wave, with maxima at the cross- 
over points. From microphotometry of 
another spectrum variable, Miss Barbara 
Middlehurst, of the Observatory of St. 
Andrews, Scotland, and Deutsch find that 
the phases and amplitudes of the velocity 
curves and line-intensity curves are strictly 
consistent with a rigid-rotator model for 
HD 124224 (A0p: P=0.52 day). 

Deutsch is making a theoretical study 
of the dynamo problem for an ideal, homo- 
geneous, spherical star. The method of 
attack involves the specification of a sta- 
tionary magnetic field which satisfies all 
approximate boundary conditions, fol- 
lowed by the solution of the equations of 
magnetohydrodynamics for the velocity 
field that this entails. This approach ap- 
pears to offer certain advantages over the 
usual method of assigning a velocity field 
and solving the equations for the associated 

Spectral Energy Distribution 

The photoelectric spectrum scanner has 
been used by Oke at the Cassegrain focus 
of the 100-inch telescope to measure the 
absolute energy distribution in the spec- 
trum of RR Lyrae as a function of phase. 
Sanford's high-dispersion spectra have 
been used to correct the scans, which have 
a resolution of only 4 A, for the effect of 
absorption lines. A comparison of the ob- 
served energy distribution of the con- 
tinuum and fluxes computed from model 
atmospheres yields accurate effective gravi- 


ties and temperatures. By combining the 
observed monochromatic light-curves and 
the effective temperatures, the change of 
radius with phase has been determined. 
This agrees roughly with the displacement 
curve obtained by integrating the radial- 
velocity curve. The comparison gives an 
estimate of the absolute radius and lumi- 
nosity of the star. The scanner has also 
been used to measure absolute energy dis- 
tributions of 16 A- and F-type giant stars 
that can be compared with RR Lyrae. 
Further studies of RR Lyrae are being 
carried out using 18 A/mm spectra ob- 
tained with exposure times of less than 15 

An analysis similar to that carried out 
for RR Lyrae is being made for the classi- 
cal cepheids v\ Aquilae, 8 Cephei, T Mono- 
cerotis, and TU Cassiopeiae. Scanner ob- 
servations for y\ Aql and h Cep (begun at 
the David Dunlap Observatory) have been 
completed by Oke. To correct the low- 
resolution scans for line absorption, high- 
dispersion spectra in the ultraviolet and 
photographic regions are being obtained 
where needed for all these stars. Absolute 
energy distributions of 14 F- and G-type 
supergiants have been measured with the 
scanner; these will serve as comparison 
stars for the cepheids. 

The scanner has been used to compare 
accurately the energy distributions of sev- 
eral late B- and early A-type stars with 
that of Vega. These stars are scattered uni- 
formly around the sky and are intended as 
standards for photoelectric spectropho- 
tometry. The stars are £ 2 Ceti, y Gemi- 
norum, a Leonis, a Coronae Borealis, a 
Lyrae, and a Pegasi. Three other stars 
used for comparison with variables can 
also be considered as standards: HD 
182487, 58 Aquilae, and 23 Cephei. All 
these stars have been compared with Vega 
or one other standard at least five times. 

A Hertzsprung-Russell diagram has 
been constructed by Oke for about 180 
late-type stars which have absolute magni- 
tudes accurate to 0.2 magnitude. The ab- 
solute magnitudes have been determined 

from accurate trigonometric and spectro- 
scopic parallaxes. Most of the stars are 
either on the main sequence or have 
evolved like stars in the older galactic 
clusters. Sixteen stars appear to be about 
twice as old as those in M 67. The main 
sequence below M v = +4 is considerably 
broader than can be explained. 

Several late-type stars which appear to 
be about 1 mag above the main sequence 
are being compared with normal main- 
sequence stars to determine whether they 
are peculiar in any way. Scanner observa- 
tions have been made by Oke to determine 
their temperatures accurately, and high- 
dispersion spectra have been obtained for 
studies of chemical abundance. 

Stars with H and K Emission Lines 

Wilson has completed and prepared for 
publication the discussion of the calibra- 
tion used for determining absolute magni- 
tudes from the emission-line widths in H 
and K. This calibration depends only 
upon observations of the sun and of the 
yellow giants in the Hyades, and fits also 
the observations of some of the M-type 
supergiants in h and x Persei. A statistical 
discussion shows that for giants and sub- 
giants a single observation of good quality 
has a probable error of about 0.3 mag, 
whereas for main-sequence stars the ac- 
curacy is less. 

In a second study, Wilson makes use of 
the observations of about 250 stars to con- 
struct a color-magnitude diagram for late- 
type stars near the sun. The H and K 
measures of giants and subgiants are sup- 
plemented by plotting trigonometric 
parallax data for stars on and near the 
main sequence whose parallaxes exceed 
0.080 second. All these points seem to fit 
well together and lead to an estimate of 
the age of the oldest local stars of about 
10 10 years. It is hoped that further observa- 
tion will increase the accuracy of this es- 

In a third investigation, Wilson has used 
these methods to locate about two dozen 


high-velocity stars in the color-magnitude 
diagram. There appears to be a difference 
between the stars of lower and higher 
space velocity. The lower-velocity group, 
having a mean space velocity of 70 km/ 
sec, tends to lie along the lower boundary 
of the general distribution for late-type 
local stars mentioned in the preceding 
paragraph. The higher-velocity stars, of 
mean space velocity 100 km/sec, however, 
tend to lie on a steeper track and to rise 
higher into the normal giant region. It is 
tentatively suggested that these differences 
are due to differences in metal content of 
the two groups, although observations are 
continuing in order to establish the effect, 
if it exists, more securely. 

Miscellaneous Studies 

The new high-dispersion grating in the 
coude spectrograph of the 100-inch tele- 
scope is being used by Oke to re-examine 
complex interstellar lines. The dispersion 
with the 114-inch camera is 1.1 A/mm, and 
the resolution of the emulsion is 0.02 A. 
The H and K lines of Ca II and the D 
lines of Na I are being observed in the 
spectra of O- and B-type stars brighter than 
fourth magnitude. Considerably more de- 
tail has been observed; the H and K lines 
in the spectrum of £ Ononis have at least 
five components, each of which has ap- 
proximately the instrumental profile. 

From some sixty dense spectrograms at 
4.5 A/mm of about forty K and M giants 
and supergiants, it now appears from 
studies by Deutsch that circumstellar cores 

can be seen at H and K in all these objects 
that are later than M0. The core first ap- 
pears at an expansion velocity of about 20 
km/sec, near the shortward edge of the 
chromospheric emission line. With ad- 
vancing type and increasing luminosity, it 
grows in strength and yields a smaller ex- 
pansion velocity. Other abundant atoms 
and ions begin to show circumstellar cores 
when the core at K exceeds about 300 mil- 
liangstroms in equivalent width. A few 
giants at M0 and K5 may show evidence 
for intermittent ejection of Ca II at speeds 
that approach the probable escape velocity 
at the reversing layer. A number of bina- 
ries involving M giants and supergiants 
are still under observation, as was noted in 
last year's report. Some of the M-type 
giants, recently discovered by Dr. Olin 
Eggen, of the Royal Greenwich Observa- 
tory, in moving clusters, show evidence of 
loss of mass; this finding is consistent with 
Sandage's conclusion that in M 67, which 
has a similar HR diagram, the stars at the 
top of the giant branch still contain more 
than half their initial hydrogen. 

A program to observe T associations has 
been started by Varsavsky. The 60-inch 
telescope is used for the UBV photoelec- 
tric observations, and the nebular spectro- 
graph of the 100-inch telescope to obtain 
spectra of the brighter members of the as- 
sociation. Photoelectric observations of 
Ophiuchi Tl in Kholopov's catalogue are 
complete; some members have also been 
observed spectroscopically. The association 
Delphini Tl will also be observed. 


The investigation of the distribution of 
gas velocities over the Orion nebula using 
multislit techniques has been completed 
and prepared for publication by Wilson 
and Munch with the assistance of Miss 
Flather and Mrs. Coffeen. 

Observations of nebular spectra for the 
purpose of fixing the wavelengths of the 
forbidden lines have been extended to the 

far ultraviolet by Bowen, using the new 
16-inch all-quartz camera of the 100-inch 
coude spectrograph. Exposures of three 
or four nights were obtained for NGC 
6572, NGC 7027, and the Orion nebula. 
New forbidden lines of N I, Ne III, Na IV, 
CI III, A III, Fe III, and Fe V were re- 
corded and measured. 



The globular and galactic clusters pro- 
vide large assemblages of stars with a wide 
range of masses, but of the same age and 
chemical composition. Large differences in 
age and composition occur from cluster to 
cluster, however. Measurements of the 
color-magnitude relationships in these 
clusters therefore yield basic information 
for the studies of stellar evolution. 

The first photoelectric results faint 
enough to reach the main sequence of a 
globular cluster were obtained by Baum 
in 1953. It was found that the main se- 
quence of M 13 lies in the subdwarf re- 
gion at the left of the ordinary main se- 
quence for stars of the kind predominat- 
ing in the solar neighborhood. Beginning 
in the spring of 1958, a joint program was 
undertaken by Dr. H. L. Johnson, of 
Lowell Observatory, Dr. W. A. Hiltner, of 
Yerkes Observatory, Sandage, and Baum 
to re-examine the same problem with 
greater accuracy and to review the inter- 

New photoelectric observations of 26 
stars in M 13 fainter than V = 17.5 mag 
were made. The data were obtained with 
the Lowell 42-inch reflector, the McDonald 
82-inch reflector, and the 200-inch reflector. 
Four different photometers of the pulse- 
counting type were used in the program. 
The photoelectric stars provided standards 
for photographic interpolation of colors 
and magnitudes of 213 additional faint 
stars in M 13. The new results confirm the 
subdwarf position ascribed earlier to the 
M 13 main sequence, but locate it with 
about three times greater accuracy. 

Included among the new photoelectric 
observations were ultraviolet measure- 
ments of 10 of the main-sequence stars in 
M 13. A mean ultraviolet excess of 0.23 
mag was found. This result is consistent 
with the ultraviolet excesses observed by 
Dr. N. G. Roman for subdwarfs in the 
general field. Line-blanketing corrections 
based upon the ultraviolet excess of M 13 
have been applied to the observed B — V 

values of the M 13 stars. After blanketing 
corrections were applied, the main se- 
quence of M 13 could be compared with 
that of the Hyades. The distance modulus 
of M 13 was thereby found to be 14.3 ±0.3 
mag. The absolute magnitude of the RR 
Lyrae stars in M 13 is therefore +0.3 ±0.3 
mag. The turn-off point of the subgiants 
from the main sequence occurs at an ab- 
solute magnitude of +4.1 ±0.3, which cor- 
responds to an age of about 10 10 years ac- 
cording to the models of Hazelgrove and 

Three-color photoelectric measures of 71 
stars, reaching as faint as ¥=22 and B = 23 
mag, have been made in the globular M 5 
by Arp. Limiting exposures with the 
//3.67 and //4.7 Ross correctors on the 
200-inch have been obtained. Final reduc- 
tion of the results is nearly complete, and 
preliminary interpretation of them indi- 
cates: (1) Three-color photoelectric meas- 
ures of noncluster members projected onto 
the same field show that the cluster is un- 
reddened. (Original work on M 5 had 
indicated, from the location of the RR 
Lyrae gap, that M 5 was reddened by a 
small amount.) (2) The RR Lyrae gap 
may be slightly wider than in M 3, but 
this preliminary indication needs to be 
checked carefully. (3) The color-magni- 
tude diagrams of M 5 and M 3 resemble 
each other more closely than any other two 
globular clusters, however, and the pre- 
liminary measures of the main sequence of 
M 5 confirm the approximate place of the 
M 3 main sequence (Sandage, 1953; John- 
son and Sandage, 1956). 

Again, pending the final results, it ap- 
pears that the globular clusters which are 
relatively rich in metals (Deutsch, Green- 
stein, previous annual reports) like M 3 
and M 5 will have main sequences which 
yield fainter absolute magnitudes for their 
RR Lyrae variables than those in a globu- 
lar cluster like M 13 (revised M 13 by 
Baum, Hiltner, Johnson, and Sandage). 

Three-color photoelectric measures of 17 


stars down to V=21 mag have been made 
by Arp in M 2. Limiting photographic 
exposures have also been obtained. Work 
is still in progress on this cluster. Very 
preliminary results for this relatively metal- 
poor cluster indicate that its main sequence 
(measured from the RR Lyrae stars) falls 
appreciably fainter than in M 5. 

A very faint star-poor cluster was dis- 
covered by Dr. S. van den Bergh, of David 
Dunlap Observatory, on National Geo- 
graphic Society-Palomar Observatory Sky 
Survey plates. Deep photographic plates 
with the 200-inch by Arp reveal that the 
cluster is probably a very poor, very distant 
globular cluster like those studied by 
E. M. Burbidge and Sandage, Abell no. 3 
and no. 4. The present cluster is not on 
Abell's list of such objects and may turn 
out to be one of the most distant of any 
of such objects. Photoelectric measures 
preliminary to obtaining a color-magni- 
tude diagram of this object had already 
been started. 

Photometry of the globular cluster NGC 
6356 (7=334°, b=+9°) has been com- 
pleted by Sandage and Wallerstein. This 
cluster, located only 11° from the direction 
of the galactic center, is the prototype of 
the group of disk clusters, first isolated by 
W. W. Morgan, which have abnormally 
strong metal lines in their integrated spec- 
tra. A photoelectric sequence reaching to 
V=19A, B = 20.3 mag was determined 
near NGC 6356 with the 200-inch tele- 
scope. Measurements of a series of plates 
show that the color-magnitude diagram 
of NGC 6356 differs significantly from that 
of typical halo clusters such as M 3, M 13, 
M 92, and M 15. There is a pronounced 
giant sequence extending from V=15.3, 
B-V = 2.2 mag, to V = 17.5, B-V=L2 
mag; but the slope of this giant branch is 
considerably less than that of a normal 
globular cluster. There appear to be a 
short horizontal branch and the beginning 
of a subgiant branch. The difference be- 
tween the horizontal branch and the giant 
branch is only 2.4 mag, as compared with 
the usual value of 3 mag. Comparison of 

the color-magnitude diagram of NGC 6356 
with other clusters has been made by fit- 
ting the horizontal branches, assuming 
Eb-\ — 0.6 mag (obtained from colors and 
spectra of foreground B and A stars) . This 
comparison shows that the giant branch 
of NGC 6356 is fainter than the giant 
branches of typical globular clusters such 
as M 13 and M 92, but brighter than the 
giant branch of the old galactic cluster 
M 67. Thus NGC 6356 is intermediate 
between M 13, which shows a metal de- 
ficiency of a factor of about 20, and M 67, 
whose stars seem to have normal metallic 
lines. These positions of the giant branches 
agree with the theoretical expectation that 
the brightness of the giant stars in globular 
clusters of nearly the same age (that is, for 
stars of nearly the same mass) is a strong 
function of the metal abundance. And, 
since the evidence is now accumulating 
that globular clusters in our own Galaxy 
do not form as homogeneous a group in 
respect to chemical composition as has pre- 
viously been believed, we must conclude 
that the brightness of the tops of the giant 
branches varies from cluster to cluster. 
This effect has consequences for the dis- 
tance scale of globular clusters as deter- 
mined from the brightest stars. 

A similar but still more abnormal result 
was obtained by Dr. Eggen and Sandage 
in their study of NGC 1783, a globular 
cluster in the Large Magellanic Cloud. 
From a two-color photoelectric sequence 
determined to V = 17.1, B = 17.9 mag, near 
NGC 1783, it was found that the giant 
branch of the color-magnitude diagram for 
NGC 1783 extended to B-V = 2.3 mag, 
and again the slope of this branch was 
considerably less steep than in halo globu- 
lar clusters such as M 3 and M 92. The 
diagram for NGC 1783 is unlike that of 
any cluster studied so far in our own Gal- 
axy. But NGC 1783 appears to be typical 
of those clusters in the Large Magellanic 
Cloud that have been classified as globular. 
The only other clusters whose diagrams 
are like that of NGC 1783 are NGC 419 


and NGC 361 studied by Arp in the Small 
Magellanic Cloud. 

These results raise doubts that the as- 
sumption of similarity of stellar objects 
can be applied from galaxy to galaxy. This 
assumption has been the basis for distance 
determination both of globular clusters in 
our Galaxy and of external galaxies. With 
the chemical composition now appearing 
to play a role in the absolute magnitudes of 
the distance indicators, an additional com- 
plication enters the picture — one that will 
be difficult to overcome in the external 

The photoelectric calibration of the ga- 
lactic cluster NGC 2158 has been com- 
pleted by Arp. It consists of a three-color 
photoelectric calibration of 29 stars to 
V=19, B = 20 mag. Photographic plates 
have been obtained. The results are being 
prepared for publication in collaboration 
with Dr. James Cuffey, of Indiana Uni- 
versity, who is also contributing observa- 
tional material of his own. The color-mag- 
nitude diagram of this cluster resembles 
that of NGC 7789 (E. M. Burbidge and 
Sandage). It is a somewhat younger clus- 
ter than M 67, of the order of age of NGC 
752. It is hoped that the study of clusters 
as rich as these will yield information on 
the return path of evolving giants and de- 
tails of the evolution in the giant region 

Three-color photoelectric measures of 30 
stars to V = 18 and B = 19 mag have been 
made by Arp in the galactic nucleus re- 
gion of NGC 6522. Also, many B and V 
photographic plates have been obtained. 
Unusually good seeing in May and June 
enabled this very difficult observing pro- 
gram (8=— 30° in the crowded nuclear 
star fields) to be essentially completed. 
This field is the one in which Baade meas- 
ured his well known results on the RR 
Lyrae stars in the galactic nucleus. The 
observational material obtained thus far 
will be analyzed to determine: (1) the 
reddening and absorption between the sun 
and the galactic center (to be derived both 
from three colors of field stars and the 

color-magnitude diagram of the globular 
cluster NGC 6522); (2) distance to the 
center (using Baade's previous values of 
the RR Lyraes with new magnitude 
scales) ; (3) the composition of the nucleus 
or kinds of stars comprising the galactic 


Sandage and Dr. Eggen discussed the 
available data on the position of subdwarfs 
in the color-magnitude diagram. All sub- 
dwarfs with adequate trigonometric paral- 
laxes (n— 0.05 second) show an ultraviolet 
excess, which, on the basis of previous 
work on line blanketing, is interpreted as 
the effect of weak Fraunhofer lines on the 
observed color indices. A line-blanketing 
theory based on solar line strengths was 
used to correct the observed B — V values 
of subdwarfs to the system of colors for 
the strong-line stars in the Hyades. The 
ultraviolet excess was used as an index of 
the line weakness. When these corrections 
to B — V were made, all the subdwarfs on 
the program moved to the Hyades main 
sequence, indicating that, to within ±0.2 
mag, the subdwarfs do not form a sequence 
separate from the normal dwarfs in the 
Mboi, log T e diagram. The conclusion is 
that the smaller metal abundance of sub- 
dwarfs does not affect their internal struc- 
ture by a large factor and therefore that 
the internal opacity of the subdwarfs must 
be predominantly due to free-free transi- 
tions of the H and He rather than to 
bound-free transitions of the heavier ele- 

The subject of line blanketing and its 
effect on the continuous radiation and 
measured colors of dwarf and subdwarf 
stars was reviewed by Melbourne. Ob- 
served energy distributions obtained with 
the spectral scanner were corrected for 
blanketing by measures of equivalent 
widths. The corrected fluxes were com- 
pared with theoretical flux distributions 
derived from model atmospheres. Reason- 
able agreement was obtained; the com- 
parisons serve to establish a set of effective 


temperatures for these stars. The UBV 
colors of these stars were computed from 
the observed energy distributions, and the 
results were compared with the observed 
colors. The effects of the line blanketing, 
with and without the hydrogen lines, on 
the UBV colors were quantitatively as- 
sessed. Line-free UBV and color-magni- 
tude main-sequence relationships were sub- 
sequently obtained. A three-dimensional 
(U — B) — (B — V) effective temperature 
relationship was established. 

Absolute Magnitude of RR Lyme Stars 

An attempt to find a new way of cali- 
brating the absolute magnitude of RR 
Lyrae stars was made by Dr. Eggen and 
Sandage using the moving-group method. 
This method utilizes the proper-motion 
and radial-velocity data for field stars to 
isolate stars moving in nearly parallel orbits 
in the solar neighborhood and therefore 
toward a common point on the celestial 
sphere. Five very high-velocity stars were 
shown to have nearly identical space vec- 
tors, and therefore apparently they form a 
kinematic group. Among the stars were 
the high proper-motion stars Groombridge 
1830 and RR Lyrae itself. The moving- 
group method gives the parallax of each 
individual star and therefore its absolute 
magnitude. The four nongiant members 
of the Groombridge 1830 group which 
satisfy the two criteria of the angle of the 
proper motion and the size of the radial 
velocity form a sequence in the color-mag- 
nitude diagram which is in the subdwarf 
region. The ultraviolet excesses of these 
stars independently identify them as sub- 
dwarfs. The absolute magnitude derived 
by the method of RR Lyrae itself is M v = 
+ 0.65 ±0.1 mag. This value differs con- 
siderably from M v = 0.00 mag usually 
adopted, but agrees well with a recent de- 
termination of Pavlovskaya using the sta- 
tistical parallax method, and it agrees 
within 0.3 mag of that determined from 
the study of M 13 discussed earlier in this 

Additional RR Lyrae stars have been 

provisionally shown to be moving with 
certain high-velocity subdwarfs. The ab- 
solute magnitude derived on the assump- 
tion of common space motion is M v = 
+ 0.6 ±0.2 mag. But clearly many more 
data on proper motions and radial veloci- 
ties must be obtained for both RR Lyrae 
stars and subdwarfs before the method can 
be fully exploited. 

Absolute Magnitude of Cepheids 

In the program to redetermine funda- 
mentally the cepheid-period luminosity re- 
lation from cepheid members of galactic 
clusters, the following further progress has 
been made by Arp : A paper on DL Cas- 
siopeiae in NGC 129 (Arp, Stephens, and 
Sandage) is in press. Work on CV Mono- 
cerotis in an anonymous cluster (Arp) has 
been completed and is about to be sub- 
mitted for publication. Arp and Sandage 
have started to repeat observations on U 
Sagittarii in M 25 to check the unpublished 
results of Dr. John Irwin, of Indiana Uni- 
versity. Photometric results on AO Canis 
Ma j oris indicate that the cepheid is not 
associated with the surrounding cluster. A 
preliminary color-magnitude diagram of 
the cluster near XZ Canis Majoris indi- 
cates that this cepheid is probably not a 
member of the cluster. The variable star 
reported by A. Briin near NGC 2355 is a 
possible member of the cluster, but the 
nature of the variable has not yet been 
established. In summary: Four cepheid 
members of galactic clusters have been 
analyzed at Mount Wilson and Palomar 
Observatories. They deviate from a linear 
period-luminosity relation by an average of 
only about ±0.15 mag, the accuracy of the 
derived luminosities. Sandage is working 
on two more, CE Cassiopeiae A and B, 
members of NGC 7790; Arp is checking 
one, U Sagittarii, and analyzing a possible 
member of NGC 2355. Two others, S 
Normae in the southern hemisphere, which 
has been completed by Irwin but not pub- 
lished, and a long-period cepheid in Puppis, 
which is in progress by Dr. R. P. Kraft and 


Ferni, exhaust the known possible cepheid 
candidates for this program. 

Nucleus of a Planetary Nebula 

The planetary nebula NGC 246 has a 
double nucleus. One member of the pair 
is extremely blue and is apparently the 
source of energy for the nebulosity. Its 
companion, at a distance of about 3 sec- 
onds of arc, is a star of average color index 
and several magnitudes fainter. 

Several attempts have been made by 
Baum in previous years to observe the com- 

ponents separately. If the fainter com- 
ponent is assumed to be a normal star on 
the main sequence, the absolute magnitude 
of the hot primary component can thereby 
be derived. Spectra of both stars have been 
obtained by Minkowski. Owing to an ap- 
parent inconsistency between the spectro- 
scopic results and the earlier photometric 
results, a renewed effort was made by 
Baum to observe the separate components 
again photoelectrically on October 11, 1958. 
The new photometric results show excel- 
lent internal consistency. 


The Local Group 

Miss Swope has continued the analysis 
of the variables on Baade's plates of the 
Draco System and of the three variable star 
fields in Andromeda. Arp, Baum, and 
Sandage each secured four photoelectric 
standards in Draco extending to the 20th 
magnitude in the photovisual. The earlier 
studies of Draco by Miss Swope had been 
dependent on transfers from the photo- 
metric sequence in M 13, which was itself 
revised in 1958. A revision of all magni- 
tudes for the color-magnitude diagram and 
the variable stars has therefore been neces- 
sary. Actually, little change was made in 
the photographic magnitudes, but the pho- 
tovisual magnitudes based on less plate 
material needed adjustment. Arp secured 
21 photoelectric standards in the variable 
star field 96 minutes from the nucleus of 
M 31 extending to 22.2 mag photovisual. 
He also measured photoelectrically a few 
more stars in the Sextans system. More and 
fainter photoelectric measures and photo- 
graphic plates extending to fainter magni- 
tudes are still required to complete the 
analysis of the system. 

A few three-color photoelectric measures 
of bright stars in M 31 and M 33 were ob- 
tained by Arp in a start on the problem 
of what are the brightest stars in these 

A five-year program of photoelectric and 
photographic photometry was begun on 

selected fields in the Large Magellanic 
Cloud at the Radcliffe and Cape Observa- 
tories in South Africa by a team from the 
Royal Greenwich Observatory in England. 
Sandage joined the project in South Africa 
for a temporary stay in October 1958 and 
worked for six months with Dr. Olin J. 
Eggen in setting up photoelectric sequences 
in two colors and obtaining photographic 
plates. Two areas of the Large Magellanic 
Cloud were surveyed with the Radcliffe 
74-inch and the Cape 18-inch reflectors. 
Photographic plates were obtained for 
studies of the cepheid variables, and of 
selected clusters in the LMC. The photo- 
electric sequences were obtained from V= 
9.6, B = 10.66 mag, to V=17.1, B = 17.9 
mag, in a field near the cluster NGC 1783, 
and from V = 8.7, B = 9.3 mag, to V=16.5, 
B = 16.4 mag, in a field near the cluster 
NGC 2004. A series of 25 plates covering 
a field of 1° were taken by J. Alexander 
with the Cape 18-inch-24-inch double re- 
fractor in B and V wavelengths to obtain 
two-color light-curves of cepheids in the 
two photoelectric regions to check the 
theoretical relation between period, color, 
and luminosity described in last year's re- 
port. Measurement of the plates has been 
begun by Dr. R. v. d. R. Woolley at 

A renewed search for dwarf galaxies has 
been started by Zwicky with the 48-inch 
schmidt telescope. One such object was 


found at R. A. 16 h 57 m 20 s and decl.-0°27' 
58". It has been checked with the Hale 
telescope and was found to contain about 
400 stars in the apparent photographic 
magnitude range from 19 to 22. The whole 
system is rather bluish, and the total lu- 
minosity of the stars fainter than m P = 22 
is very feeble. The apparent diameter of 
the system is about 11 minutes of arc. Es- 
timating the distance at half a million 
light-years, the absolute magnitude of the 
whole system would be between M p = — 4 
and —5, and the absolute diameter of the 
order of 1500 light-years. This dwarf gal- 
axy, which has the appearance of a uni- 
formly populated open cluster, is the in- 
trinsically faintest galaxy known so far. 

Rotation and Dispersion of Stellar 

It has long been known that different 
galaxies show absorption lines of different 
widths. For instance, NGC 221 (M 32) has 
relatively narrow lines, NGC 224 (M 31) 
rather wide lines. The quantitative de- 
termination of the dispersion of stellar ve- 
locities, which is the source of the widen- 
ing of the lines, is difficult since there are 
usually no suitable lines, unaffected by 
blending with neighboring lines, whose 
contours can be determined with some de- 
gree of accuracy. For NGC 221 and NGC 
224 it was possible for Minkowski to de- 
termine the velocity dispersion by compar- 
ing the spectra of the galaxies with stellar 
spectra whose lines were widened artifi- 
cially (Carnegie Institution Year Book 53, 
p. 26, 1954). Excessive exposure times ac- 
quired even with the smallest dispersion of 
the coude spectrograph at the 200-inch tele- 
scope made it necessary for Minkowski to 
transfer the observations of other galaxies 
to the prime-focus spectrograph, where a 
dispersion of 85 A/mm is being used. Since 
it is not feasible to obtain with this spectro- 
graph suitably modified stellar spectra for 
comparison, the quantitative evaluation of 
the spectrograms presents certain difficul- 
ties which have not yet been entirely over- 
come. The qualitative inspection of the 

spectrograms, however, has already led to 
the important result that the velocity dis- 
persion is not correlated with the absolute 
magnitude and that the width of the lines 
does not furnish a criterion for absolute 
magnitude. The most crucial case is an 
anonymous E0 galaxy, 7.3 minutes north 
preceding NGC 4486, M pg -15.2, whose 
lines are much wider than those of NGC 
221, E2, M p —15.4, and almost as wide as 
those of NGC 4472, M v -21.3. This result 
leads to the conclusion that spherical gal- 
axies do not necessarily have similar mass 
distributions — only if this were true, a cor- 
relation between luminosity and velocity 
distribution would have to be expected. 

The investigation of the rotation in 
NGC 3115 is essentially completed by 
Minkowski. In the central part out to 
about 2 seconds of arc from the center, the 
rotation is high ; the radial velocity here in- 
creases by 42 km/sec per second of arc. 
The increase of velocity slows down almost 
abruptly to about 25 km/sec per second of 
arc at about 2 seconds from the center. The 
rotational velocity reaches a maximum of 
250 km/sec at a distance of 16 seconds, de- 
creases to a broad minimum of 190 km/sec 
at a distance of 40 seconds, and increases 
again as far out as the observations could 
be made, reaching a value of 325 km/sec 
at 90 seconds. The determination of the 
mass density following from the new re- 
sults, to be undertaken in cooperation with 
Dr. J. H. Oort, of the University of Leiden, 
is not yet finished. 

The results for the rotation in NGC 3115 
are in some ways reminiscent of those ob- 
tained much earlier by H. W. Babcock for 
NGC 224, particularly in the high rota- 
tional velocity in the immediate neighbor- 
hood of the center. On the still unmeasured 
spectrograms of other galaxies this phe- 
nomenon can be observed in several cases 
where the nuclear region is large enough 
to be resolved and the absorption lines are 
narrow enough to show relatively small 
velocity differences. The most striking one 
is NGC 4111, long known for its outstand- 
ingly high rotation in the central part, 



which exhibits a very rapid change to 
rather slow rotation in the outer parts. It 
seems possible that high rotation in the 
nuclear region is a general feature that 
escapes observation when the nuclear re- 
gion is too small to be resolved. 

Stellar Populations in Ellipticals 

The program initially reported on last 
year consists in measuring near-by ellip- 
tical galaxies in eight different colors rang- 
ing from ultraviolet to infrared and de- 
termining the mixture of stars necessary 
to fit the results obtained. It has been 
found by Baum that there is a definite di- 
vision between elliptical galaxies brighter 
than —15 absolute magnitude and those 
fainter than —15. All the large bright el- 
lipticals have intrinsic color indices around 
B — V— 0.9 mag, whereas the dwarf ellip- 
ticals have intrinsic color indices around 
B-V = 0.6 mag. 

The results of the analysis for stellar con- 
tent imply that the dwarf ellipticals are 
predominantly composed of metal-poor 
stars (Population II), whereas large el- 
lipticals are predominantly composed of 
metal-rich stars (Population I). Both the 
populations are old, because no new stars 
are being formed in elliptical galaxies. 

These results have important evolution- 
ary implications. In our own Galaxy, the 
stars of Population II have very high ran- 
dom velocities and were formed during a 
brief epoch of perhaps half a billion years 
at the very beginning of the history of the 
Galaxy. In a dwarf galaxy, star formation 
has apparently not progressed beyond an 
initial Population II epoch. In large ellip- 
ticals, star formation apparently continued 
considerably past the Population II epoch 
but came to a halt before the present day. 
It evidently continued long enough for 
heavy elements to be produced by nucleo- 
genesis in extremely massive stars and to be 
returned to the interstellar medium in time 
to become ingredients for a substantial frac- 
tion of the stars formed subsequently. The 
only galaxies in which star formation has 
continued until the present are the mod- 

erately large systems which also have a 
high degree of flattening (an oblateness 
ratio greater than 3 to 1) due to rotation; 
these are the spirals. In short, we can draw 
two rough dividing lines — one at an ab- 
solute magnitude of — 15 and the other at 
an oblateness ratio of 3 to 1. This fact sug- 
gests that the evolutionary history of a gal- 
axy is largely determined by two initial 
conditions: its mass and its angular mo- 

The Distribution of Luminosity 

The distributions of luminosity in four 
elliptical galaxies of the Virgo Cluster were 
observed photoelectrically during the re- 
port year by Baum. Some 20 objects have 
been similarly observed in various clusters 
in previous years. The general purpose of 
these observations is to provide surface- 
brightness profiles from which effective 
angular sizes can be determined and from 
which the relative distances of various 
clusters can thereby be inferred. The im- 
mediate purpose of the photoelectric ob- 
servations in Virgo is to provide photoelec- 
tric calibrations for a larger number of 
luminosity distribution measurements to be 
done photographically in Virgo. The pho- 
tographic plates for this work were ob- 
tained at the 48-inch schmidt telescope, and 
a special processing was used to obtain 
finer grain and better photometric uni- 
formity than are obtained in standard proc- 
essing. All the observational material will 
be turned over to Dr. Martha Hazen Liller, 
of the University of Michigan. She will de- 
rive isophotal contours of a number of 
Virgo galaxies by measuring the plates 
with the recording isophotometer of the 
University of Michigan. The photoelectric 
data will be used for calibrating the re- 
sulting isophotes. 

Velocities of Galaxies 

By the procedure described in previous 
annual reports, the photoelectric measure- 
ments of redshifts and magnitudes of re- 
mote galaxies have been continued by 
Baum. Each galaxy is observed in several 


colors, and the radiant energy is plotted as 
a function of wavelength. The displace- 
ment of the resulting curve from its normal 
position provides a measure of the apparent 
velocity of recession. As was reported last 
year, redshifts as large as 0.4 of the velocity 
of light have been reached by this method. 
The largest redshift observed during the 
current report year was 0.33 of the velocity 
of light, or 100,000 km/sec. This was the 
mean value obtained for four galaxies in 
CL 0024 + 1654, a remote cluster of galaxies 
in the southern galactic hemisphere. 

The purpose of Baum's program is to 
extend the relationship between redshift 
and distance (the distance inferred from 
the apparent magnitude and the apparent 
angular diameter) so as to distinguish be- 
tween various possible cosmological models. 


The search for supernovae has continued 
at several American and European observa- 
tories under the general direction of 
Zwicky with the partial aid of the U. S. 
National Science Foundation and the Swiss 
Federal Science Foundation. During the 
report year, four supernovae were found at 
Palomar, as follows: (1) Supernova in 
NGC 1350. Found by Gates on 18-inch 
schmidt photographs of January 1959. (2) 
Supernova in NGC 4921, a member of the 
Coma cluster of galaxies. When first ob- 
served by Humason with the 48-inch in 
June 1959, its estimated photographic mag- 
nitude was 18.5. Since then it has slowly 
decreased in brightness. (3) Supernova in 
a faint anonymous Sc galaxy at R. A. 13 h 
9"M, decl. +3° 36:4 (1960). First observa- 
tion of the supernova (estimated magni- 
tude 14) was made by Humason with the 
48-inch in June 1959. The redshift of the 
galaxy is unknown, although an attempt 
will be made to determine it in the next 
observing season in order to settle the ques- 
tion of possible membership in the Virgo 
cluster. One spectrum of this supernova in 
the blue region was obtained by Greenstein 
at the 200-inch on July 2-3. This object 
showed several strong features. A very 

strong dip centered at A3800 separates an 
apparent emission maximum at A3680, 200 
A wide, from more diffuse bands in the 
blue-green. The spectrum seems to be that 
of a type I supernova. (4) Supernova in 
NGC 7331. This was first observed by 
Humason with the 48-inch schmidt in June 
1959 and had an estimated magnitude of 
13. The spectrum of this supernova was 
observed more extensively in the ultra- 
violet and blue regions by Greenstein, 
June 30-31 through July 2-3. Very wide 
spectra (0.8 mm) were obtained at 180 
A/mm. An additional spectrum was ob- 
tained by Preston with the 100-inch on July 
5-6. Double interstellar K absorption lines 
were separated by about 800 km/sec. The 
spectrum extends far into the ultraviolet. 
It has very weak broad emission (or ab- 
sorption) features. Relatively sharp emis- 
sions at AA3385 and 3560 are visible. Shal- 
low dips in the spectra, which may be 
broad absorption lines or gaps between 
broad emission lines, are visible at AA3790, 
4250, and 4750. The spectrum may be that 
of a type I supernova at an unusual stage 
of development, in which all features are 
very weak as compared with those on 
Minkowski's spectra. Whether this object 
represents a new type of supernova spec- 
trum, or a hitherto unknown phase, or may 
even develop toward a type II spectrum, is 
not yet clear. 

Three-color photoelectric magnitude 
standards have been set up around the two 
brightest supernovae, four stars having 
been measured near (3) and five near (4). 
Photoelectric measures of the supernovae 
themselves have been obtained by Arp, and 
photographic records of these supernovae, 
from discovery to date, have been obtained 
by Humason, W. Miller, Arp, Perek, and 

Clusters of Galaxies 

The first volume of the catalogue of gal- 
axies and clusters of galaxies which is being 
prepared by Zwicky with the collaboration 
of Herzog and Dr. P. Wild is nearly ready 
for publication and will contain data on 


about 9000 galaxies and 2000 clusters. This 
project is supported in part by the Office 
of Naval Research. 

The large-scale distribution of clusters of 
galaxies has been studied by Zwicky in the 
belt R. A. 8 b to 16 h and decl. -5° to +20°, 
about 1500 distant, very distant, and ex- 
tremely distant clusters being recorded. 
Wherever the number of these various 
classes of clusters per unit area is small, it 
is found that a near-by or medium-distant 
large cluster of galaxies covers the field. 
This is additional proof that intergalactic 
absorbing matter within the large clusters 
of galaxies dims the light coming from 
more distant galaxies. Within large areas 
not covered by any rich near-by or me- 
dium-distant clusters of galaxies, the dis- 
tributions of distant, of very distant, and 
of extremely distant clusters of galaxies are 
individually and collectively of a remark- 
able uniformity and of a complete ran- 
domness, in confirmation of the conclu- 
sion previously announced that no cluster- 
ing of clusters of galaxies occurs. 

A comprehensive program has been 
started by Zwicky for the investigation of 
the velocity dispersion of the members of 
interconnected multiple galaxies and of 
galaxies in groups and in clusters of vari- 
ous sizes. From these investigations it is 
hoped to derive pertinent data on the prop- 
erties and origins of all these systems. Val- 
ues for M/D, where M is the total mass of 
the group or cluster and D is its absolute 
distance, can also be derived. This deter- 
mination of M/D is independent of any 

assumptions about the absolute luminosity 
of various characteristic types of stars, such 
as have been chosen in the past for the 
establishment of distance scales (variables 
of various kinds, novae, supernovae, and 
brightest stars in galaxies) . 

The spectral types of the various classes 
of structurally distinct galaxies in groups 
and clusters are also being investigated. 
Finally, new information has been gained 
on the spectra of the luminous intergalac- 
tic bridges. So far about 100 spectra have 
been obtained with the prime-focus spectro- 
graph of the Hale telescope. The spectra 
of the galaxies interconnected by luminous 
filaments and luminous clouds appear par- 
ticularly interesting and challenging. 

At the same time as the spectroscopic in- 
vestigations, the spatial distribution of the 
galaxies in groups and clusters is being 
studied by Zwicky, using for every cluster 
three exposures in the blue and three in 
the red. All these investigations have con- 
firmed the assumption that most galaxies 
must be considered as being associated with 
groups and clusters and that these are virtu- 
ally space fillers. Also, much new evidence 
has been obtained on the partial segrega- 
tion within clusters of galaxies of different 
brightness and different structural types. 

Direct photographs have been obtained 
by Zwicky of a number of widely sepa- 
rated galaxies which have been found to 
be interconnected, generally by luminous, 
bluish intergalactic formations of stars. 
Some of these are remarkably filamentary 
in structure. 


As was mentioned in the introduction, 
the Pyrex mirrors in the coelostats of the 
two solar towers have been replaced by 
mirrors of fused quartz. This improve- 
ment has almost eliminated the shift in 
focus that formerly occurred in these in- 
struments during the first hour after being 
exposed to sunlight. Others include a 
photoelectric guider for the 150-foot tower 
and extensive modifications of the spectro- 
heliograph of the 60-foot tower. 

To facilitate the investigation of the de- 
tailed patterns of the magnetic field in ac- 
tive solar regions, and their changes with 
time, H. W. Babcock and Howard have 
developed an alternative scanning and re- 
cording system for the magnetograph of 
the 150-foot tower telescope. This equip- 
ment permits precise automatic scanning 
of a region some 4 minutes of arc square 
with a resolution of 5 seconds of arc in a 
time of 15 minutes. Conformal recording is 


accomplished by photographing a cathode- 
ray tube on which the magnetic pattern is 
displayed with intensity modulation in five 
distinct levels, changing abruptly at 5, 10, 
20, and 50 gauss. The magnetic polarity is 
indicated by stretching the recording spot 
into a short line that slants to the right or 
left. These records are more informative 
and pictorially more attractive than the 
older type. The new technique permits 
sequential scanning of active regions at the 
rate of 3 or 4 per hour for extended inter- 
vals. It is expected that the rate of varia- 
tion of the field patterns, particularly in 
connection with solar flares, will be sig- 

The Cassegrain spectrum scanner is be- 
ing modified and completely remounted 
for Oke. The new mounting will make it 
much easier to find objects and will pro- 
vide offset guiding. 

The coude scanner, referred to in last 
year's report, has been built and installed 
in the coude spectrograph at Mount Wilson 
under the direction of Oke and Wilson. 
In this instrument the grating and photo- 
multiplier cells remain fixed; scanning of 
the spectrum is accomplished by rotating a 
long double 45° offset prism about an axis 
passing through the scanning slit. Ap- 
proximately 60 A can be scanned at any 
one grating position. The monitored light 
signal is a part of the spectrum adjacent to 
that being scanned. The mechanical and 
optical parts perform very satisfactorily. 
The detection, amplification, and recording 
of the scan and monitor signals can be 
accomplished separately at present. The 
problem of obtaining the ratio of the two 
signals at the telescope has not been com- 
pletely solved. 

A Moseley X-Y plotter and curve-fol- 
lower has been added to the micropho- 
tometer at the California Institute of 
Technology by Oke. This instrument 
automatically translates densities into in- 
tensities continuously by means of calibra- 
tion curves. The intensity tracing is made 
in the usual way with a strip chart 

Parker has studied the spectrophoto- 
metric properties of the Palomar wedge 
and coude spectrographs. Only vanish- 
ingly small systematic errors were found 
to exist as a function of wavelength in the 
visual region. Intercomparison of Mount 
Wilson and Palomar spectra of the same 
stars is under way. Tests of the micropho- 
tometer used at the California Institute, 
either as a transmission-measuring device 
or as an automatic linear device (employ- 
ing a Moseley curve-follower), have proved 

A new vacuum tank which will permit 
aluminizing the 72 -inch mirror of the 48- 
inch schmidt and the secondary mirrors of 
the 200-inch telescope has been designed by 
Rule and constructed for the Palomar Ob- 
servatory. During the coating process the 
mirror or other surface is held in a vertical 
position. This will permit the application 
of various over-coats and nonreflection 
coats which must be evaporated from a 
"boat" rather than from a filament. 

In order to expedite the measurement of 
spectrum plates, one of the Observatories' 
8-inch measuring engines has been fitted 
with an automatic digital "print out" re- 
cording system. This eliminates the tedi- 
ous manual recording of the data and the 
errors that may arise from this procedure. 
The integrated digital recording system 
consists of (1) a conventional measuring 
engine; (2) a Coleman six-decade digit- 
izer, coupled to the screw of the measuring 
engine through a gear train; (3) a Clary 
printer, which prints the settings on a tape. 
The range of measurement is 210 mm with 
a tolerance of 1 micron. 

A camera for rapidly and easily making 
finding charts for the 100-inch and 200- 
inch telescopes from the National Geo- 
graphic-Palomar Observatory Sky Atlas 
was designed by Miller and Van Devender 
and constructed in the Santa Barbara Street 
shop. The camera makes use of a 4 by 5 
Land camera back and Polaroid Land film 
packets. It is designed for a fixed magni- 
fication of 5, which brings the scale of the 
Sky Atlas (focal length = 10 feet) up to ap- 



proximately that of the larger instruments 
(100-inch, /=42 feet; 200-inch, /=55 feet). 
The area to be photographed is very 
quickly located by positioning the Sky At- 
las print under a transparent plastic plate 
with an aperture corresponding to the area 
covered by the photograph. 

The ruling of 10 large blazed gratings 
was undertaken during the year by Swan- 
son under the supervision of H. W. Bab- 
cock. One of them, 6 by 10 inches in size 
with 900 grooves per millimeter, has proved 
useful in the coude spectrograph of the 
100-inch telescope. It provides a dispersion 
of 4.5 A/mm with the 32-inch camera and 
1.1 A/mm with the 9-foot camera. Large 
coude gratings were also supplied to the 
Mount Stromlo and Dominion Astrophysi- 
cal Observatories. 

The first successful attempts at these Ob- 
servatories to photograph nebulae and gal- 
axies in color were made public in April 
of this year. Begun more than two years 
ago by Miller, the program took advantage 
of the exceptional optical speed of the 48- 
inch schmidt (F/2.5) and the 200-inch 
Hale (F/3.3) telescopes on Palomar Moun- 
tain, and the unique sensitivity of Super 
Anscochrome color film. 

Preliminary tests at the 48-inch schmidt 
indicated that many nebulae and galaxies 
were within the reach of the two large 

instruments. Extensive laboratory tests 
were required to determine the perform- 
ance characteristics of the color film with 
the long exposure times necessary for faint 
astronomical objects. Unfortunately the 
three color-sensitive emulsions comprising 
the film have different reciprocity charac- 
teristics, so that a serious loss of color bal- 
ance occurs as exposure times are increased. 
Methods for the accurate determination 
and control of this color shift were worked 
out in cooperation with local Ansco en- 
gineers. As a result, the picture films ob- 
tained at the telescope could be corrected 
for the unavoidable loss of color accuracy, 
in so far as the color characteristics of the 
film permit and within the fundamental 
limitations of any three-color process. 

The first results of this color program 
were displayed at the California Institute 
of Technology and the Carnegie Institu- 
tion of Washington in the form of large 
color transparencies, and reproductions 
were published in the National Geographic 
Magazine for May 1959 and in Life for 
April 27, 1959. 

A summary of the techniques worked 
out for accurate color control of astro- 
nomical photographs will appear shortly in 
the Publications of the Astronomical So- 
ciety of the Pacific. 


As has been the practice in recent years, 
the Observatories have invited a number 
of guest investigators to make use of their 
observational facilities. The following pro- 
grams have been carried through during 
the report year. 

Dr. George O. Abell, of the University 
of California at Los Angeles, has been in- 
vestigating about three dozen rich clusters 
of galaxies for the purpose of determining 
the bright end of the galaxian luminosity 
function. With the 48-inch schmidt tele- 
scope he has obtained series of extrafocal 
photographs of the clusters on both blue- 
and yellow-sensitive emulsions. Magni- 

tudes are derived from measures of the 
extrafocal galaxian images by a technique 
developed by Dr. Abell and Mr. Dimitri 
Mihalas. The photographic technique has 
been calibrated, and stellar magnitude se- 
quences are being set up in the cluster 
fields with photoelectric observations made 
at the 100-inch telescope. 

The reductions are now complete for 
the Coma cluster of galaxies. The lu- 
minosity function shows a preliminary 
maximum about 2 mag below the bright 
end. One must count through about 150 
galaxies to reach the maximum. Following 
it is a trough about 1.5 mag wide, after 


which the luminosity function increases 
rapidly with increasing magnitude. If the 
feature proves to be common to the lu- 
minosity functions of the other rich clus- 
ters, it should be a valuable distance cri- 
terion for galaxian clusters. 

Radial velocities for the clusters in this 
program for which velocities are not 
known are also being measured with the 
Newtonian spectrograph of the 100-inch 
telescope. In particular, velocities have 
been measured for six clusters that appear 
to belong to a second-order grouping of 
clusters near a = \6 h , h— +30°. As yet, not 
enough data are available to establish an 
accurate velocity dispersion among the 
clusters in the group, but the spread of 
velocities obtained so far is from 8700 to 
13,000 km/sec, a range that is not incom- 
patible with the assumption of gravita- 
tional interactions among the clusters. 

On March 8-9, a flare was observed by 
chance on AD Leonis, a type dM4e star 
that was being observed in the course of 
establishing magnitude sequences in clus- 
ter fields. Three-color observations of the 
flare indicated B-V=0.0, and U-B = 
— 1.4 mag, for the flare itself. Its maximum 
luminosity was at least 1.47 X 10 3 ° ergs/sec, 
and the total energy liberated was at least 
3X10 32 ergs. 

During the report year, a study was 
nearly completed by Dr. Abell of 86 plane- 
tary nebulae that were found on the Na- 
tional Geographic Society-Palomar Ob- 
servatory Sky Survey. Distances have been 
derived by the method of Shklovsky (Russ. 
Astron. ]., 33, 222, 1956), and absolute mag- 
nitudes have been derived for the central 
stars. Some of the stars appear to have 
dimensions of white dwarfs. 

During the current year Dr. Dinsmore 
Alter made 99 lunar photographs with the 
60-inch reflector for the study of various 
lunar features. The technique was identical 
with that of the previous year except for 
trial use of Kodak IV-N plates on the full 
moon. The brightness at this phase made 
the slower plates with their fine grain prac- 

ticable. Their greater contrast proved to be 
useful in obtaining full-phase detail. 

In collaboration with Dr. A. G. Wilson, 
of the Rand Corporation, Dr. Alter made 
26 lunar spectrograms with the X spectro- 
graph of the 60-inch telescope in a pre- 
liminary study of possible methods of in- 
vestigation of gaseous phenomena on the 

During the fall, Dr. Guillermo Haro, of 
the Tacubaya and Tonantzintla Observa- 
tories, and Dr. W. J. Luyten, of the Uni- 
versity of Minnesota Observatory, spent 
two dark-of-the-moon observing periods on 
Palomar Mountain using the 48-inch 
schmidt telescope in their program of faint 
blue stars in high galactic latitude. They 
took some 50 plates in the region surround- 
ing the south galactic pole following 
Haro's three-image method with exposures 
through ultraviolet, yellow, and blue filters, 
respectively. Preliminary examination of 
these plates has resulted in the discovery 
of some 5000 blue stars brighter than the 
19th magnitude. 

The region of the sodium D lines in the 
spectra of 16 Mira variables of type Me 
taken with the 32-inch camera at the coude 
focus of the 100-inch telescope was studied 
by Dr. Philip C. Keenan, of the Perkins 
Observatory. On plates of scale 15 A/mm 
it was found possible to separate the inter- 
stellar D lines from the stellar lines when 
the peculiar velocity of the star exceeded 
about 30 km/sec. Of the eight observed 
variables with V— 30 km/sec, six showed 
interstellar D lines for which displacements 
and equivalent widths could be measured. 
Only one low-latitude variable, X Gemini, 
failed to show the expected interstellar 
components. X Gem thus appears to lie 
in a direction where the concentration of 
sodium is low. 

For X Monoceros, at galactic latitude 
— 1°, comparison with four near-by stars 
of types B or A allowed Dr. Keenan to 
estimate roughly the visual absolute mag- 
nitude of the variable near maximum light. 
A plot of interstellar intensity against 


m—M gave M»= —1.1. From the appear- 
ance of its spectrum, also, X Mon seems 
to be one of the least luminous of the Mira 

During the past year, a technique has 
been improved by Dr. R. B. Leighton, of 
the Physics Department of the California 
Institute, whereby solar magnetic fields and 
line-of -sight velocities can be mapped pho- 
tographically. In a preliminary test of these 
techniques rather strong magnetic fields 
(up to 200 gauss) were observed through- 
out plage regions, the field pattern being 
so strikingly correlated with the bright 
Ca II emission as to suggest a close causal 
relation between them. 

Calibrated magnetograms of the solar 
disk were obtained on 94 days at the 150- 
foot solar tower in the period from July 1, 
1958, to December 31, 1958, by Dr. William 
Livingston, of the University of California. 
The circumstances attending these observa- 
tions continued to be the same as those re- 
ported in the previous year. The sensitivity 
averaged 5 gauss per interval of the dis- 
played trace. Microfilmed copies of these 
observations were distributed to interested 
persons as a contribution to the Interna- 
tional Geophysical Year. In addition to the 
disk magnetic records, polar traces with 
high sensitivity were made on 95 days. Be- 
cause of the increased sensitivity, these ob- 
servations demand individual attention in 
order to fix the zero point, scale, and assess 
the general quality, and are thus not suit- 
able for circulation to other groups as raw 
data. These polar observations are being 
studied by Dr. H. D. Babcock and Dr. 
Livingston in order to determine fluctua- 
tions in the general magnetic field of the 
sun. Cragg and Hickox contributed to the 
above program. 

During August, experimental observa- 
tions of the solar spectrum were obtained 
by Dr. Livingston with an image orthicon 
as the receiver at the focus of the 75-foot 
spectrograph (150-foot tower). The resolu- 
tion of this tube closely matches the linear 
resolution of the spectrograph in the fifth 
order near A5000. As employed in these 

tests, the light-transfer characteristic of the 
image orthicon is approximately the same 
as that of a fine-grain photographic emul- 
sion. In common with the emulsion, the 
image orthicon receives and responds to 
the light of each spectral element simul- 
taneously. Thus seeing, transparency 
changes, and intrinsic fluctuations in the 
solar and telluric line strengths act only 
to smear the spectral features in a sym- 
metrical way. The image orthicon is found 
to have an important photometric advan- 
tage over the emulsion in that the sensi- 
tivity can be calibrated in the exact area 
that is exposed. It was found that equiva- 
lent widths down to about 20 X 10 -6 A 
could be measured. The equipment proved 
too unreliable to undertake a systematic 
program of observations at the time. An 
improved model is currently under con- 
struction at Kitt Peak National Observa- 
tory, Tucson, Arizona. 

From April through October 1958, pho- 
toelectric observations of members of a re- 
stricted class of early B giants have been 
obtained with the Palomar 20-inch by Dr. 
Roger Lynds, of the University of Cali- 
fornia at Berkeley. These observations 
were made in an effort to establish con- 
stancy or variability in light. A restricted 
list of stars was prepared which included 
stars whose MK type and luminosity class 
agree to within half a spectral type sub- 
division and half a luminosity class with 
those of known (3 Canis Majoris stars. Only 
stars north of h= — 10° with m<7.0 mag 
were included. The list includes 50 stars 
and is 70 per cent complete for m<.6.5 
mag. It is probably better than 50 per cent 
complete for ra<7.0 mag, and thus repre- 
sents a good sample. 

The list of 50 stars includes seven known 
(3 Canis Majoris stars and ten spectroscopic 
binaries, of which one is a known eclipsing 
binary and another a known elliptical 
variable. In addition, two stars (Be stars) 
are known to vary in light in an irregular 
way and twelve have definitely variable 
radial velocities. 

The reduction of the observations is still 


in progress, but the following results are in- 
dicated: (a) four new probable 3 CMa 
stars were identified; {b) three possible 3 
CMa stars were found; (c) two new ellipti- 
cal variables were found; and (d) eight 
stars with definitely variable light but of ir- 
regular or undetermined type were found. 

It can be concluded in a rough way from 
existing data, and in particular from an in- 
vestigation by M. F. Walker (Astrophys. 
/., 57, 227, 1952), that the 3 CMa phenome- 
non is restricted to a fairly small range of 
MK type. Dr. Lynd's present investigation 
shows that in this range of MK types ap- 
proximately 25 per cent of the stars are 3 
CMa stars. 

The program of observations of the solar 
spectrum with the Snow telescope has been 
continued by the McMath-Hulbert Ob- 
servatory, the observer being Mr. Thomas 
K. Jones. As in earlier years, the programs 
of the Snow telescope have been closely in- 
tegrated with the programs at Lake Ange- 
lus. During the time covered by this report, 
the program of the Snow telescope was 
almost exclusively the production of obser- 
vations for the derivation of abundances, in 
support of the McGregor solar tower pro- 
gram. The Snow telescope observations 
have been especially valuable because of the 
increased ultraviolet transparency of the 
atmosphere over Mount Wilson in com- 
parison with that over Lake Angelus. In 
addition to the direct recording of spectral 
line profiles, drift curves of the sun at se- 
lected positions in the ultraviolet have been 
obtained. These are essential parts of the 
data required for the determination of 
chemical abundances in the sun's atmos- 

Dr. D. H. McNamara, of Brigham 
Young University, has completed the in- 
vestigation of RZ Scuti. During the cur- 
rent year he has studied the velocity curve 
outside of eclipse and has accounted for the 
remarkable distortions present in it by as- 
suming moving streams of gas that distort 
the spectral lines to give fictitious veloci- 
ties. It has been possible to reconstruct the 
true velocity curve of this system by cor- 

recting the strongly distorted helium line 
for this effect. 

Spectrograms of the eclipsing system 
AW Pegasi have been obtained by Dr. Mc- 
Namara in the blue, yellow, and red re- 
gions. Previous investigations of the star 
have indicated that the spectrum is com- 
posite but have used so low dispersion that 
no well defined velocity curve could be ob- 
tained for either component of the blended 
lines. The present plates show Ha and the 
sodium D lines to be distinctly double, 
which should make possible the determi- 
nation of a velocity curve for both com- 

Rotation velocity curves (V e sin i) based 
on the assumption that line widths are due 
to rotation were determined by Dr. McNa- 
mara for all the brighter 3 CMa stars. 
Values ranging from to 40 km/sec were 
found. Some stars were observed with spec- 
tra resembling the 3 CMa stars and with 
rotational velocities in the same to 40 km/ 
sec range that do not show the variable 
radial velocities characteristic of the 3 CMa 
stars. From a statistical point of view, ap- 
proximately one-half of the B stars of the 
proper spectral types and luminosity class 
and V e sin / less than 40 km/sec were found 
to belong to the 3 CMa group of variable 
stars. This is approximately the number to 
be expected if it is assumed that pulsation 
occurs in all B stars of proper spectral type 
and luminosity class below a certain critical 
velocity of about 40 km/sec. Thus, the B 
stars that resemble the 3 CMa stars but 
apparently have no variable radial velocity 
in a short period are those for which the 
V e sin i is less than 40 km/sec because 
of a small sin i factor. Another indication 
that rotation may play an important role is 
the fact that only those stars with relatively 
large rotational velocities exhibit beat phe- 
nomena in their velocity curves. 

Dr. W. W. Morgan, of Yerkes Observa- 
tory, continued his classification of the 
forms of galaxies. In the period Septem- 
ber 10 to October 10, 1958, the classification 
of all galaxies of the Shapley-Ames Cata- 
logue brighter than 13.0 mag and north of 


— 25° declination was completed. A num- 
ber of somewhat fainter systems were also 
classified. The 1958 work depended exclu- 
sively on the National Geographic Society- 
Palomar Observatory Sky Survey plates. 

The system on which the galaxies were 
classified was that developed by Dr. Mor- 
gan and described in Pubs. Astron. Soc. Pa- 
cific, 70, 364-391, 1958. This material will 
be used to prepare a catalogue of galaxies 
classified on the revised Yerkes system. 

Dr. L. Plaut, of the Kapteyn Laboratory 
at Groningen, Netherlands, has taken 71 
plates with the 48-inch schmidt camera for 
the Groningen program on faint variable 
stars. With the X spectrograph (4-inch 
camera) of the 60-inch reflector, 41 plates 
were taken by Dr. Plaut for a cooperative 
program with Dr. A. Blaauw on radial ve- 
locities of early-type stars in the associa- 
tions Lacerta and Cepheus III. 

Mass-luminosity and H-R diagrams have 
been constructed by Dr. Daniel M. Popper, 
of the University of California at Los An- 
geles, for the components of eclipsing bi- 
naries lying near the main sequence. When 
only results of high precision are employed, 
these stars conform to the same relations as 
near-by wide visual binaries and single 
stars. The cosmic scatter in the relation- 
ships also appears to be no greater for the 
close binaries than for single or wide dou- 
ble stars. The need to confine the discus- 
sion to eclipsing systems with well observed 
properties is stressed. Uncertainty in the 
temperature scale for the hotter stars is the 
greatest cause of error in the comparisons, 
since the temperature enters directly into 
the evaluation of luminosities for the 
eclipsing stars, whereas it does not enter 
directly in evaluating the luminosities of 
the near-by stars. 

Results of the revision by Popper of the 
properties of RX Herculis on the basis of 
improved data appeared in the Astrophys. 
]., 129, 659-667, 1959. The rediscussion of 
RS Canes Venatici was nearly ready for 
publication when photoelectric observations 
in 1959 showed that the light at total eclipse 
had changed by 0.3 mag since earlier obser- 

vations in 1956 without corresponding 
changes in the light outside eclipse. Ap- 
parently the subgiant component is varia- 
ble irregularly both in time and over the 
surface of the star. This is not a system in 
which the subgiant component nearly fills 
the critical zero-velocity surface. RZ Erid- 
ani, announced years ago as a metallic- 
line eclipsing binary, is found not to be- 
long to this class of stars, showing instead 
a true composite spectrum. 

Dr. A. T. Purgathofer, of the Vienna 
University Observatory, made an extensive 
series of photoelectric measurements on the 
following clusters: NGC 581, 1807, 1817, 
1857, 2215, 2301, 2323, 2324, 2360, 6871, 6883, 
IC 4996, and fields near 25 Cygnus and 29 
b3 Cyg. In addition, both photographs and 
photoelectric measurements were made of 
the following clusters: NGC 1502, 1664, 
1907, 2506, and 5053. Photoelectric meas- 
ures were made in the UBV system to pro- 
vide calibration scales for the construction 
of color-magnitude diagrams of these sys- 

A program to study the color variation 
in galaxies photoelectrically was started by 
Dr. Morton S. Roberts, of the University 
of California at Berkeley. Measurements 
in three colors were made along the major 
and minor axes of the elliptical galaxy 
M 32. The central region of M 32 is defi- 
nitely redder, as shown by both B — V and 
U — B colors. The gradient of the color var- 
iation along the two axes is different, be- 
ing greater along the minor axis. This last 
result implies a spectral-luminosity func- 
tion for M 32 which varies with distance 
from the equatorial region of the galaxy, 
or some reddening matter within the 
galaxy which is concentrated toward the 
equator, or a combination of both effects. 
Further observation on other elliptical neb- 
ulae may make it possible to distinguish 
between true ellipticity and orientation 

Blue and infrared plates of face-on spi- 
rals were obtained with the 60-inch tele- 
scope to study the structure of the arms and 
nucleus in the light of these two colors. The 


latter plate -filter combination avoids the 
emission lines from H II regions. The "hot 
spots" in the nucleus noted by Morgan ap- 
pear to be giant H II regions. 

Additional photographic work with the 
48-inch schmidt telescope was carried out 
by Dr. Roberts in a search for loose galactic 
clusters in the Case LF regions for which 
spectral types have been published. 

Drs. John Rogerson, Lyman Spitzer, and 
John Bahng, of the Princeton University 
Observatory, brought to Mount Wilson a 
pulse-counting photometer and tested it for 
high-dispersion spectrophotometry in the 
coude spectrograph of the 100-inch tele- 
scope. "Seeing compensation" was achieved 
with two phototubes, one measuring a nar- 
row spectral band (about 0.12 A) which 
could be scanned over about 145 A, the 
other monitoring a 12-A band. Although 
losses at the entrance slit and at various re- 
flections decreased the over-all sensitivity, 
the seeing fluctuation, as was evidenced in 
the scatter of 1-minute counts, was reduced 
from about 8 per cent in each channel sepa- 
rately to about 0.6 per cent in the ratio of 
the two channels. The instrument was used 
for several preliminary observations of in- 
terstellar lines. In particular, measurements 
on the D lines gave equivalent widths 
agreeing systematically with earlier Mount 
Wilson photographic determinations to 
about 2 per cent. 

A spectrographic investigation of the ra- 
diation of faint meteors or the faint stages 
of the radiation of bright meteors was un- 
dertaken by Dr. John A. Russell, of the 
University of Southern California. On the 
nights of August 11, 12, and 13, 1958, forty- 
two 103a-E films were exposed in the 18- 
inch schmidt telescope, equipped with its 
objective prism, for a total actual exposure 
time of 6 h 07 m . The spectrum of one Perseid 
meteor was recorded, in which about 30 
well defined lines and an equal number of 
questionable lines were measured. The ra- 
diation can be attributed, with one notable 
exception, to transitions of low excitation 
potential in atoms of Fe, Na, Mg, Ca, and 
possibly Ni and Al; there is no convincing 

evidence that any radiation from ionized 
atoms is present. The exception mentioned 
is the strongest feature of the spectrum, an 
unresolved band whose wavelength range 
of 6370 to 6630 A agrees with that of the 
first positive group of bands of molecular 
nitrogen. {See G. Herzberg, Molecular 
Spectra and Molecular Structure, vol. 1, 
Spectra of Diatomic Molecules, fig. 8, 
p. 32, 1950.) Not only is this feature the 
strongest, but also it is visible about three 
times as far up the trail (toward the 
radiant) from the plate edge as any other 
spectral line. The spectrum obtained adds 
to the knowledge of meteoric radiation at 
a stage when the meteoric matter is not 
yet incandescent and all the photographic 
radiation is coming from atmospheric 

In October 1958, Mars was observed with 
the 200-inch telescope by Dr. William M. 
Sinton, of Lowell Observatory, who used 
an infrared spectrometer equipped with a 
liquid-nitrogen-cooled lead sulfide cell. The 
primary aim of his program was to confirm 
the provisional 1956 discovery of Martian 
infrared bands that are produced probably 
by organic molecules. The large scale pro- 
vided by the 200-inch made it possible to 
establish that the bands occur chiefly in the 
dark areas of Mars, and only very weakly 
in the bright areas. All together, three dis- 
tinct absorptions near 3.5 microns were 
found. Bands in approximately the same 
positions are found in reflection spectra of 
terrestrial plants. In plants these bands are 
produced by organic molecules. One of 
them is generally produced by carbohy- 
drates. Hence, these observations furnish 
strong evidence for plant life in the dark 
areas on Mars. Other than these, and bands 
of carbon dioxide previously found by 
Dr. G. P. Kuiper, no Martian absorption 
bands were observed in the range 1 to 4.2 

In order to learn more about the gas 
streams in 3 Lyrae and their evolutionary 
significance, simultaneous spectrographic 
and photoelectric observations were made 
by Dr. Otto Struve, of the University of 


California, on Mount Wilson, and by 
Mr. David Wood and Dr. Merle Walker 
at Lick Observatory. The results indicate 
that the light-curve was broader during the 
principal eclipse of June 23-27, 1958, than 
during the following eclipse of July 6-10, 
1958. The shell absorption lines of He I 
3820, 4026, 4472, and 4713 were consider- 
ably stronger during the first eclipse than 
during the second. The gas stream that 
produces the violet "satellite lines" was also 
more pronounced during the first eclipse. 
Evidently this gas stream, which originates 
in the B8 component near the Lagrangian 
point Li, is irregular in density. When it 
contains more than the average number of 
absorbing He I atoms, the increased density 
is observed a day later as an increased 
amount of absorption in the He I shell 
lines that are produced at low levels in the 
expanding shell of the system. This in- 
crease in the amount of gas produces a de- 
crease in the total brightness and is also 
responsible for a marked change in the 
velocity of expansion of the shell. 

A comparative study of the spectra of all 
3 Canis Majoris variables observed by Dr. 
Struve at Mount Wilson (in which Dr. 
Abhyankar and several graduate students 
participated) resulted in a modified H-R 
diagram for these stars, from which they 
conclude that the period of the pulsation 
does not define their locations uniquely. 
Differences in chemical composition that 
cannot be explained in terms of simple evo- 
lutionary processes are probably involved, 
and perhaps also differences in age. 

The radial velocity measurements of the 
components of 48 visual double stars were 
prepared for publication by Dr. Struve. 
Several peculiar spectroscopic binaries were 
observed in different spectral regions in 
order to study the behavior of gas streams 
and the apsidal motions. 

Dr. George Wallerstein, of the Univer- 
sity of California at Berkeley, has contin- 
ued several spectroscopic investigations. 
The spectrum of VY Canis Majoris was 
photographed from A4000 to A8800. Al- 

though the star was considerably brighter 
than a year earlier, the spectrum appeared 
to be largely unchanged. Observations of 
this remarkable star will be continued. 

Radial velocity observations of the he- 
lium-rich star o Orionis E show that its ve- 
locity is not variable and is close to that of 
star AB, thus confirming that star E is a 
member of the a Orionis system. 

The composite spectrum (about B8V + 
KOII) of star a in M 41 offers an oppor- 
tunity to obtain the mass ratio of a red 
giant and a star that is at the top of the 
main sequence of a galactic cluster. This 
provides a test of the hypothesis that the 
cluster giants have masses slightly greater 
than those of the brightest main-sequence 
stars. Nine spectra at a scale of 20 A/mm 
and 10 A/mm were obtained between De- 
cember 1958 and March 1959. Measure- 
ments of six of these, well distributed in 
time, show no change in velocity, indicat- 
ing that the period must be rather long. 
Observations will be continued. 

Dr. Wallerstein is obtaining spectra at 
15 A/mm in the visual region of high-ve- 
locity G dwarfs in order to make abun- 
dance studies by comparison with the sun. 
Stars with a wide variety of galactic orbits 
and with ultraviolet excesses from to 0.2 
mag are included in the program. 

Mars was observed photographically 
with the 60-inch by Dr. A. G. Wilson, of 
the Rand Corporation, and Dr. R. S. Rich- 
ardson, of the Griffith Observatory, on a 
total of 32 nights throughout its 1958 pe- 
riod of close approach, beginning with 
August 3, 105 days before opposition (No- 
vember 16), and terminating on March 22, 
126 days after opposition. The principal 
object of the program was to study in 
greater detail the phenomenon of the "blue 
clearing." For this purpose, Mars was 
photographed several times each night of 
observation in both blue (III-O or Process) 
and in yellow light (IV-F + GG II filter). 
On selected nights comparative spectra of 
Mars and the moon were made in the 
A3600 to A4700 region to study the blue 


haze continuum. A total of 33 spectra and 
165 direct plates were secured with more 
than 1400 direct images of Mars. 

Strong or weak clearings were observed 
on 17 of the 32 nights on which observa- 
tions were made, clearings being observed 
as early as September 3, 74 days before op- 
position, and as late as March 13, 117 days 
after opposition. These results stand in con- 
tradiction to the previously held idea that 
clearings occur only within a few days of 
opposition. Further, clearings were found 
not to be planet-wide phenomena, as had 
previously been held. They could appear 
and fade in only a few hours, and an ex- 
ample of a clearing rotating onto the visible 

disk was recorded on November 30. The 
spectral studies did not definitively confirm 
the minima at A4250 and A4495 in the spec- 
tra of the blue haze continuum which were 
reported by A. G. Wilson in 1956. Con- 
tinued suspicion of the presence of such 
dips, however, calls for the development of 
more sensitive techniques for future oppo- 
sitions. Numbers of clouds were photo- 
graphed on the blue plates, including the 
W-shaped cloud observed in 1954. Near 
opposition during periods of clearing, it 
was found that the blue clouds on the sun- 
set limb were located over the desert re- 
gions and those on the sunrise limb over 
the maria. 


Dr. Alexander Pogo retired as Librarian 
and Editor on June 30, 1959. He joined the 
staff of the Division of Historical Research 
of the Institution on July 1, 1929, to assist 
Dr. Sarton in his researches on the history 
of science. For the next 21 years he carried 
out extensive investigations on sixteenth- 
century science, Egyptian astronomy, and 
Maya chronology and its relationship to 
Christian chronology. During most of this 
period he served as associate editor of his. 

In 1950 Dr. Pogo transferred to the 
Mount Wilson Observatory. His back- 
ground in the history of science has en- 
abled him to strengthen historical collec- 
tions in the Hale room of the Mount Wil- 
son Library. Similarly, he has been of 
great assistance in cataloguing and obtain- 
ing supplementary material for the collec- 
tion of early books on astronomy purchased 
a few years ago for the library in Robinson 
Hall. These two libraries now provide the 
Observatories with one of the most com- 
plete collections of early astronomical 
books available in the western United 
States. Dr. Pogo's unusual knowledge of 
practically all European languages has been 
of great help in handling the Observa- 
tories' correspondence and in keeping the 
staff in touch with the recent Russian litera- 
ture in astronomy. 

Dr. Guido Munch spent most of the year 
as Guggenheim Fellow and Fulbright 
Scholar at the Max Planck Institute in 
Munich and the Observatory at Leiden. 

Research Division 
Staff Members 
Halton C. Arp 

Horace W. Babcock, Assistant Director 
William A. Baum 
Ira S. Bowen, Director 
Armin J. Deutsch 
Jesse L. Greenstein 
Fred Hoyle 

Rudolph L. Minkowski 
Guido Munch 
J. Beverley Oke 
Allan R. Sandage 
Olin C. Wilson 
Fritz Zwicky 

Research Associate 
Edwin E. Salpeter 

Senior Research Fellows 
Geoffrey R. Burbidge 
H. Lawrence Heifer 
Satoshi Matsushima 
Bernard Y. Mills 

Carnegie Research Fellows 
Robert F. Howard 

George W. Preston » 

Carlos M. Varsavsky , , 


Research Fellows 
Walter K. Bonsack 
E. Margaret Burbidge 
Alercio M. Gomes 
Alois T. Purgathofer 
S. Saito 

Gerhard Traving 
George Wallerstein 

Research Assistants 
Sally Bonsack 
Frank J. Brueckel 
Sylvia Burd 
Mary F. Cofleen 
Thomas A. Cragg 
Dorothy S. Deutsch 
Edith Flather * 
Emil R. Herzog 
Joseph O. Hickox 
A. Louise Lowen 
Mildred Matthews 
Cynthia Stephens 2 
Henrietta H. Swope 
Robert L. Wildey 

Student Observers 

William G. Melbourne 
Robert A. R. Parker 

Editor and Librarian 
Alexander Pogo 3 


William C. Miller 

Instrument Design and Construction 

Lawrence E. Blakee, Electronic Technician 
Kenneth J. Cochran, Draftsman 
Floyd E. Day, Optician 
Kenneth E. DuHuff, Machinist 
Robert D. Georgen, Machinist 
Don O. Hendrix, Superintendent of Optical 

1 Resigned October 17, 1958. 

2 Resigned March 15, 1959. 

3 Retired June 30, 1959. 

Melvin W. Johnson, Optician 
Stuart L. Roberts, Instrument Maker 
Bruce Rule, Project Engineer 
Oscar Swanson, Instrument Maker y 
David Thomas, Electronic Technician 
Russell R. Van Devender, Jr., Designer and 
Superintendent of Instrument Shop 

Maintenance and Operation 
Mount Wilson Observatory and Offices 
Audrey A. Acrea, Stewardess 
Paul F. Barnhart, Truck Driver 
Ashel N. Beebe, Superintendent of Con- 
Wilma J. Berkebile, Secretary 
Hugh T. Couch, Carpenter 
Helen S. Czaplicki, Editorial Typist 
Stewart F. Frederick, Janitor 
Eugene L. Hancock, Night Assistant 
Mark D. Henderson, Gardener 
Anne McConnell, Administrative Assistant 
Leah M. Mutschler, Stenographer and Tele- 
phone Operator 
Bula H. Nation, Stewardess 
Alfred H. Olmstead, Night Assistant 
Arnold T. Ratzlaff, Night Assistant 
John E. Shirey, Janitor and Relief Engineer 
Benjamin B. Traxler, Superintendent 

Palomar Observatory and Robinson Labora- 
Fred Anderson, Machinist 
Dorothea Davis, Administrative Secretary 
Eleanor G. Ellison, Secretary and Librarian 
Arlis R. Grant, Stewardess 
Leslie S. Grant, Relief Night Assistant and 

Byron S. Hill, Superintendent 
Charles E. Kearns, Night Assistant 
J. Luz Lara, Apprentice Laborer 
Harley C. Marshall, Office Manager 
George W. Pettit, Janitor 
Robert E. Seares, Night Assistant 
William C. Van Hook, Electrician and As- 
sistant Superintendent 
Gus Weber, Assistant Mechanic 


Arp, Halton C. Southern hemisphere photom- 
etry. II. Photometric measures of bright stars. 
Astron. ]., 63, 118-127 (1958). 

Arp, Halton C. Stars in the direction of the 
galactic center. Astron. J., 64, 33-34 (1959). 

Arp, Halton C. The Hertzsprung-Russell dia- 

gram. Astron. Soc. Pacific Leaflet 360, 8 pp. 
(1959). In Handbuch der Physik, 51, 73-133, 
Berlin, Springer-Verlag, 1958. 
Arp, Halton C. Cepheids in galactic clusters. III. 
EV Set in NGC 6664. Astrophys. J., 128, 
166-173 (1958). 


Arp, Halton C. Reduction of photoelectric ob- 
servations by an electronic computer. Astro- 
phys. J., 129, 507-512 (1959). 

Arp, Halton C, and William G. Melbourne. 
Color-magnitude diagram for the cluster 
M 22. Astron. ]., 64, 28-33 (1959). 

Baade, Walter. Problems in the determination 
of the distance of galaxies. Astron. J., 63, 
207-210 (1958). 

Baade, Walter. Galaxies and their stellar popu- 
lations. A review of the present state of 
affairs. Ricerche astronomiche, 5, 3-21 

Baade, Walter. Physical variable stars and stellar 
populations. Ricerche astronomiche, 5, 165- 
167 (1958). 

Baade, Walter. The population of the galactic 
nucleus and the evidence for the presence of 
an old population pervading the whole disk 
of our Galaxy. Ricerche astronomiche, 5, 
303-315 (1958). 

Babcock, Harold D. See Babcock, Horace W. 

Babcock, Horace W. Magnetic fields of the 
A-type stars. Astrophys. J., 128, 228-258 

Babcock, Horace W. Deformable optical ele- 
ments with feedback. /. Opt. Soc. Am., 48, 
500 (1958). 

Babcock, Horace W. Stellar magnetic fields. 
Intern. Astron. Union Symposium 6, 161- 
165 (1958). 

Babcock, Howard W., and Harold D. Babcock. 
Photospheric magnetic fields. Intern. Astron. 
Union Symposium 6, 238-247 (1958). 

Baum, William A. A method of searching for 
clusters of galaxies beyond present limits. 
Pubs. Astron. Soc. Pacific, 70, 450-451 

Baum, William A. Population inferences from 
star counts, surface brightness, and colors. 
Pubs. Astron. Soc. Pacific, 71, 106 (1959). 

Baum, William A., and J. D. Strong. Basic 
optical considerations in the design of a solar 
furnace. Solar Energy in /. Solar Energy 
Sci. and Eng., 2, 37^15 (1958). 

Baum, William A. See also Butler, C. T.; Tuve, 
M. A. 

Burbidge, E. Margaret, and G. R. Burbidge. Ro- 
tation and internal motions in NGC 5128. 
Astrophys. J., 129, 271-281 (1959). 

Burbidge, E. Margaret, and G. R. Burbidge. 
Stellar evolution. In Handbuch der Physi\, 
51, 134-295, Berlin, Springer-Verlag, 1958. 

Burbidge, E. Margaret, and A. Sandage. The 
color-magnitude diagram for the galactic 
cluster NGC 7789. Astrophys. J., 128, 174- 
184 (1958). 

Burbidge, E. Margaret. See also Burbidge, Geof- 
frey R. 

Burbidge, Geoffrey R. The origin of cosmic rays. 
Nuoi'o cimento, 8, 403-420 (1958). 

Burbidge, Geoffrey R., E. M. Burbidge, and 
W. A. Fowler. Electromagnetic phenomena 
in cosmical physics. Intern. Astron. Union 
Symposium 6, 222-236 (1958). 

Burbidge, Geoffrey R. See also Burbidge, E. 

Butler, C. T., and W. A. Baum. The measure- 
ment of integrated extraterrestrial radiation 
by means of a satellite-borne photometer. 
Jet Propulsion Lab. Publ. 151 (1959). 

Connelley, Mary, and A. Sandage. Photoelectric 
observations of RS Ophiuchi. Pubs. Astron. 
Soc. Pacific, 70, 600-602 (1958). 

Cragg, Thomas A. Photograph of a "white 
light" flare. Pubs. Astron. Soc. Pacific, 71, 
56-57 (1959). 

Cragg, Thomas A. A light curve of the recent 
outburst of RS Ophiuchi. Pubs. Astron. Soc. 
Pacific, 71, 47-48 (1959). 

Cragg, Thomas A. See also Nicholson, Seth B. 

Deutsch, Armin J. Magnetic fields of stars. In 
Handbuch der Physi\, 51, 689-722, Berlin, 
Springer-Verlag, 1958. 

Deutsch, Armin J. Harmonic analysis of the 
periodic spectrum variables. Intern. Astron. 
Union Symposium 6, 209-221 (1958). 

Feige, Jacques. A search for underluminous hot 
stars. Astrophys. J., 128, 267-272 (1958). 

Feige, Jacques. An atlas of identification charts 
for 113 blue stars. (Abstract.) Astrophys. J., 
129, 600 (1959). 

Flather, Edith. See Osterbrock, Donald E. 

Ford, W. K., Jr. See Tuve, M. A. 

Fowler, W. A. See Burbidge, Geoffrey R. 

Greenstein, Jesse L. High-resolution spectra of 
comet Mrkos (1957d). Astrophys. J., 128, 
106-113 (1958). 

Greenstein, Jesse L. Dying stars. Sci. American, 
200, 46-53 (1959). 

Greenstein, Jesse L. See also Heifer, H. L.; 
Tifft, W. G. 

Hall, J. S. See Tuve, M. A. 

Heifer, H. L. A 21-cm survey around the 
Pleiades. (Abstract.) Pubs. Astron. Soc. 
Pacific, 70, 455 (1958). 

Heifer, H. L., and H. E. Tatel. A 21-cm survey 
around the Pleiades. Astrophys. ]., 129, 565- 
582 (1959). 

Heifer, H. L., G. Wallerstein, and J. L. Green- 
stein. Abundances in some Population II K 
giants. Astrophys. J., 129, 700-719 (1959). 

Heifer, H. L. See also Wallerstein, G. 

Hoyle, Fred, and Olin C. Wilson. Some theo- 
retical aspects of H and K emission in late- 
type stars. Astrophys. ]., 128, 604-615 (1958). 

Joy, Alfred H. Mira Ceti. Astron. Soc. Pacific 
Leaflet 358, 8 pp. (1959). 

Joy, Alfred H. A spectrographic observation of a 
flare of UV Ceti. Pubs. Astron. Soc. Pacific, 
70, 505-506 (1958). 

Melbourne, William G. See Arp, Halton C. 


Merrill, Paul W. Nebular lines in the spectrum 
of AG Pegasi. Astrophys. J., 129, 44-49 

Merrill, Paul W. The story of AG Pegasi. Sfy 
and Telescope, 18, 490-492 (1959). 

Merrill, Paul W. See also Sanford, Roscoe F. 

Miller, William C. Color in the universe. Eng. 
and Sci, 22 (no. 7), 26-31 (1959). 

Miller, William C. First color portraits of the 
heavens. Natl. Geograph. Mag., 115 (no. 5), 
670-679 (1959). 

Minkowski, Rudolph L. Cygnus loop and some 
related nebulosities. Revs. Modern Phys., 30, 
1048-1052 (1958). 

Minkowski, Rudolph, and D. Osterbrock. Inter- 
steller matter in elliptical nebulae. Astro- 
phys. J., 129, 583-595 (1959). 

Munch, Guido. Internal motions in the Orion 
Nebula. Revs. Modern Phys., 30, 1035-1041 

Munch, Guido. Kinematics of the filaments in 
the Crab Nebula. Revs. Modern Phys., 30, 
1042-1046 (1958). 

Munch, Guido, and Luis Munch. On the distance 
of the Cassiopeia radio source. Astrophys. ]., 
129, 854-856 (1959). 

Munch, Guido, and Arne Slettebak. A new 
O-type subdwarf. Astrophys. J., 129, 852- 
854 (1959). 

Munch, Luis. See Munch, Guido. 

Nicholson, Seth B., Th. Cragg, and others. Sum- 
mary of Mount Wilson magnetic observa- 
tions of sunspots for March, 1958, to De- 
cember, 1958. Pubs. Astron. Soc. Pacific, 70, 
415-420, 508-512, 611-615; 71, 58-63, 173- 
178 (1958). 

Osterbrock, Donald E. A study of two comet 
tails. Astrophys. J., 128, 95-105 (1958). 

Osterbrock, Donald E. Two dense nebulae. 
Pubs. Astron. Soc. Pacific, 70, 399-403 

Osterbrock, Donald E. Comets' tails at large dis- 
tances from the sun. (Abstract.) Pubs. 
Astron. Soc. Pacific, 70, 457-459 (1958). 

Osterbrock, Donald E. Parallel filamentary 
structure in diffuse nebulae. Pubs. Astron. 
Soc. Pacific, 71, 23-27 (1959). 

Osterbrock, Donald E., and Edith Flather. Elec- 
tron densities in the Orion Nebula. II. (Ab- 
stract.) Astron. J., 63, 310 (1958). 

Osterbrock, Donald E., and E. Flather. Electron 
densities in the Orion Nebula. II. Astrophys. 
J., 129, 26-43 (1959). 

Osterbrock, Donald E. See also Minkowski, R. 

Pettit, Edison. The visual magnitudes of double 
stars. Astron. J., 63, 324-328 (1958). 

Sandage, Allan. Cepheids in galactic clusters. I. 
CF Cass in NGC 7790. Astrophys. J., 128, 
150-160 (1958). 

Sandage, Allan. On the intrinsic colors of RR 

Lyrae stars in M 3. Astrophys. J., 129, 596- 
599 (1959). 

Sandage, Allan. A supernova in NGC 23. Pubs. 
Astron. Soc. Pacific, 71, 162 (1959). 

Sandage, Allan. The color-magnitude diagrams 
of galactic and globular clusters and their 
interpretation as age groups. Ricerche as- 
tronomiche, 5, 41-56 (1958). 

Sandage, Allan. Luminosity function of galactic 
clusters, globular clusters, and elliptical gal- 
axies. Ricerche astronorniche, 5, 75-90 

Sandage, Allan. NGC 2264 and M 8 as examples 
of very young stellar associations which are 
still partly in the stage of Kelvin contraction. 
Ricerche astronorniche, 5, 149-158 (1958). 

Sandage, Allan. The stars within 15 parsecs of 
the sun. Ricerche astronorniche, 5, 287-298 

Sandage, Allan. See also Burbidge, E. Margaret; 
Connelley, Mary. 

Sanford, Roscoe F., and P. W. Merrill. Mount 
Wilson spectrograms of AB Aurigae. Pubs. 
Astron. Soc. Pacific, 70, 602-604 (1958). 

Schmidt, Maarten. Rate of star formation. Astro- 
phys. J., i29, 243-258 (1959). 

Slettebak, Arne. See Munch, Guido. 

Smith, Harlan J. See Tifft, W. G. 

Strong, J. D. See Baum, William A. 

Tatel, H. E. See Heifer, H. L. 

Tifft, William G. A system of three-color pho- 
tometry with applications to the galactic 
structure in Cygnus. Astron. ]., 63, \27-\45 

Tifft, William G. Classical cepheids in galactic 
clusters. Astrophys. J., 129, 241 (1959). 

Tifft, William G. Photometry of extragalactic 
nebulae. (Abstract.) Pubs. Astron. Soc. 
Pacific, 70, 463-464 (1958). 

Tifft, William G., and J. L. Greenstein. MWC 
603, a high-velocity symbiotic star. In 
&oiles a raies d' emission, pp. 449-451, Liege, 
Institut d'Astrophysique, 1958. 

Tifft, William G., and Harlan J. Smith. T. Sex- 
tantis, an RR Lyrae star of type "c." Astro- 
phys. ]., 127, 591-605 (1958). 

Tuve, M. A., W. K. Ford, Jr., J. S. Hall, and 
W. A. Baum. Results of preliminary tests 
of cascaded image converters. Pubs. Astron. 
Soc. Pacific, 70, 592-594 (1958). 

Wallerstein, George. Note on the Population II 
cepheid region in the color-magnitude dia- 
gram of globular clusters. Astrophys. ]., 128, 
141-142 (1958). 

Wallerstein, George. Effective temperatures, 
radii, masses, and pulsation properties of 
the Population II cepheids, M 5 No. 42 and 
W Virginis. Astrophys. ]., 129, 356-361 

Wallerstein, George. The spectrum of the ir- 


regular variable VY Canis Majoris. Pubs. 
Astron. Soc. Pacific, 70, 479-484 (1958). 

Wallerstein, George. Sharp lines in the spectrum 
of Nova RS Ophiuchi. Pubs. Astron. Soc. 
Pacific, 70, 537-543 (1958). 

Wallerstein, George, and H. L. Heifer. Abun- 
dances in G dwarf stars. I. A comparison of 
two stars in the Hyades with the sun. 
Astrophys. J., 129, 347-355 (1959). 

Wallerstein, George, and H. L. Heifer. Abun- 
dances in G dwarf stars. II. The high-velocity 
star 85 Pegasi. Astrophys. J., 129, 720-723 

Wallerstein, George. See also Heifer, H. L. 

Weidemann, Volker. The atmosphere of the 
white dwarf Van Maanen 2. Pubs. Astron. 
Soc. Pad fie, 70, 466-A67 (1958). 

Wilson, Olin C. Internal kinematics of the 
planetary nebulae. Revs. Modern Phys., 30, 
1025-1029 (1958). 

Wilson, Olin C. See also Hoyle, Fred. 

Wilson, Ralph E. Roscoe Frank Sanford, 1883— 
1958. Pubs. Astron. Soc. Pacific, 70, 360-363 

Zwicky, Fritz. Is Newton's law of gravitation 
really universal? Astronautics, 4, 18, 19, 74, 
and 76 (1959). 

Zwicky, Fritz. A new cosmological theory. Eng. 
and Sci., 22 (no. 5), 20-21 (1959). 

Zwicky, Fritz. Future prospects of jet propulsion. 
In Jet Propulsion Engines, vol. 12 of High 
Speed Aerodynamics and Jet Propulsion, 
edited by O. E. Lancaster, Princeton Uni- 
versity Press, Section L, pp. 757-785, 1959. 

Zwicky, Fritz. Supernovae. In Handbuch der 
Physi\, 51, 766-785, Berlin, Springer-Verlag, 

Zwicky, Fritz. Multiple galaxies. In Handbuch 
der Physi\, 53, 373-389, Berlin, Springer- 
Verlag, 1959. 

Zwicky, Fritz. Clusters of galaxies. In Handbuch 
der Physil^, 53, 390-415, Berlin, Springer- 
Verlag, 1959. 

Zwicky, Fritz. Morphologishe Forschung. Win- 
terthur (Switzerland) Kommissionsverlag, 

Zwicky, Fritz. Les premiers spectres obtenus de 
la metiere brillante intergalactique. L'As- 
tronomie, Bulletin de la societe astronomique 
de Prance, 72, pp. 285-290 (Juillet-Aout, 

Zwicky, Fritz. Nuclear goblins and flare stars. 
Pubs. Astron. Soc. Pacific, 70, 506-508 


Washington, District of Columbia PHILIP H. ABELSON, Director 



Introduction 81 

Experimentation at High Pressures and Tem- 
peratures 82 

Pyrope 83 

Quartz-coesite transition 87 

Thallium transition 88 

Petrology 89 

Eclogites 89 

Pseudoleucite in a tinguaite from the Bear- 
paw Mountains, Montana 94 

Cordierites 96 

Natural cordierite-bearing rocks 96 

Anhydrous cordierites and the system 

MgO-Al 2 3 -Si0 2 98 

Cordierite-water system 100 

The effect of varying oxygen content in a 

natural rock sequence 104 

Garnets 106 

The join grossularite-pyrope at atmos- 
pheric pressure 107 


system 109 

Garnet-cordierite parageneses 112 

The system forsterite-diopside-silica-al- 

bite 113 

Feldspars 118 

Alkali amphiboles 121 

Biotites 126 

Biotites on the join phlogopite (KMg 3 - 
AlSi 3 O 10 (OH),)-annite (KFe 3 AlSi 3 - 

O 10 (OH) 2 ) .: 127 

"Ferri-annite" 132 

Mullite-H 2 system 132 

Spinels 134 

Tourmaline 137 

Ore Minerals 138 

The Ni-S system 139 

The Fe-Ni-S system: The phase relations 
between pyrite and vaesite in the pres- 
ence of excess sulfur 142 

The Fe-As-S system: The upper stability 
curve of the pyrite-arsenopyrite as- 
semblage 145 

The Ni-As-S system 148 

Natural assemblages in the Co-Ni-Fe-As 

system 153 

Application of the pyrrhotite-pyrite equi- 
librium relations 155 

Application of the pyrrhotite X-ray de- 
terminative curve to natural pyrrho- 

tites 155 

Temperatures of formation of coexisting 
pyrrhotite, sphalerite, and pyrite from 

Highland Surprise Mine, Idaho 156 

Thermal calculations pertaining to ore de- 
position 157 

Measurement of vapor pressures of sul- 
fides 160 

Differential thermal analysis 161 

Ore solutions: The system ZnS-H 2 S- 
FLO 163 


Iron Meteorites 167 

The system Fe-Ni-S 168 

The Ages of Roc\s and Minerals 170 

Mineral ages in the Maryland piedmont. . 171 
Mineral ages in the southern Appalach- 
ians 174 

Comparison of ages of minerals from peg- 
matites and enclosing granite at Cut- 
ler, Ontario 176 

Preliminary study of the age relationships 

of some Alpine rocks 177 

Acknowledgments 178 

Radioactive Fallout Particularly from the 

Russian October Series 178 

Organic Geochemistry: Kerogen 181 

Mechanisms of formation of kerogen. ... 182 

Components of ancient kerogen 184 

Testing of Steel 186 

Optical and Electrical Properties of Silicates . 187 

Crystallography 189 

Symmetry of magnetic structures 189 

Tables of magnetic space groups 189 

Alternative derivation of the magnetic 

space groups 190 

The distribution of spins in magnetic 
crystal structures of the sodium chlo- 
ride type 192 

Crystallographic study of pyroxenes 192 

The structural relations between diop- 

side, clinoenstatite, and pigeonite ... 193 
The structural relations among three 
polymorphs of MgSiO s — enstatite, pro- 

toenstatite, and clinoenstatite 197 

The crystal structure of hexagonal CaAl 2 - 

Si 2 O s 198 

The crystal structure of Fe 3 Se 4 199 

Single-crystal studies of Cu 9 S 5 -Cu 5 FeS 4 

solid solutions 201 

Natural bornite 201 

Synthetic bornite 201 

Conclusions 202 

Bornite-digenite mix-crystals 202 

Computation of Diffraction Effects of Short- 
Range Ordering in "Layered" Se- 
quences 203 

The basic computation 203 

Diffuse diffraction from the row-by-row 

model 205 

Comparison of computed and observed 

transforms 205 

Averaging within and across arrays 207 

Averaging across arrays 208 

Acknowledgments 208 

Miscellaneous Administration 209 

Penologists' Club 209 

Seminars 209 

Lectures 210 

Summary of Published Wor\ 211 

Bibliography 217 

References Cited 218 

Personnel 222 

Carnegie Institution of Washington Year Bool{ 58, 1958-1959 


Science presents a changing pattern of 
discovery and emphasis. Fads and fashions 
arise and disappear, but some broad trends 
continue for a generation or more. A 
major development is the increasing ap- 
plication of physical science to the observa- 
tional disciplines. The great expanding 
effort in medical and biological research 
is essentially an example of this. Perhaps 
even more spectacular in earth sciences has 
been the stimulus of the International Geo- 
physical Year and the current exploration 
of what is somewhat erroneously termed 
"outer space." In the effort to understand 
the solid earth physical science is also being 
applied at a heightened tempo. This type 
of study of the solid earth is the research 
program of the Geophysical Laboratory. 

The effort is certain to yield results of 
enormous long-term practical consequence 
and abundant findings of philosophic 
significance. Some of the highlights of 
the Geophysical Laboratory's contributions 
during the past year are outlined in the 
following paragraphs; the work is de- 
scribed in detail in later sections. 

Boyd and England are exploring min- 
eralogical changes occurring under pres- 
sures up to 50,000 bars, which is equivalent 
to a depth of about 150 kilometers. They 
have shown that orthopyroxene and min- 
erals rich in alumina that are commonly 
found in rocks of the earth's crust will 
react, under the conditions of temperature 
and pressure present in the mantle, to form 
the magnesian garnet pyrope. Their data 
provide additional evidence for the hy- 
potheses that one of the major minerals in 
the mantle is pyrope-rich garnet and that 
the Mohorovicic discontinuity marks a 
transition from basalt to pyrope-bearing 

These garnet-pyroxene mineral assem- 
blages called eclogites have been consid- 
ered a potential source of basalt, which at 
least in oceanic areas must come from 
within the mantle. By proper choice of 

experimental conditions Yoder and Tilley 
have converted eclogite into basalt, pyrox- 
enite, or hornblendite. By partial melting 
of eclogite either of two major types of 
basalt may be generated. 

Zies and Chayes have completed the 
first successful combined micrometric and 
chemical analysis of a pseudoleucite, sup- 
plemented by detailed chemical studies of 
the associated groundmass. Although pres- 
ent in very different amounts, the princi- 
pal phases common to groundmass and 
pseudoleucite have the same composition 
in both. In this instance the pseudoleucite 
appears to have been generated in large 
part by a base exchange at magmatic tem- 

Cordierite is an important diagnostic 
mineral in determining conditions of meta- 
morphism. Schreyer, Schairer, and Yoder 
have outlined the stability field for com- 
positions close to the magnesian end mem- 
ber. A further series of petrological experi- 
ments includes studies of pyroxenes by 
Schairer and Morimoto; garnets, Yoder, 
Schairer, and Chinner; biotites, Wones 
and Eugster; amphiboles, Ernst; spinels, 
Turnock; and feldspars, Orville. 

Tilton and Davis have been studying 
geological processes using a large number 
of nuclear clocks. The versatility of the 
original five dating methods, especially 
those employing K-A and Rb-Sr, has been 
extended through measurement of two or 
more coexisting phases in a given rock 
assemblage. The conditions prevalent dur- 
ing orogeny or metamorphism can affect 
biotite, muscovite, and potassium feldspar 
clocks differently but in systematic ways. 
The respective measured ages in several 
areas seem to correspond to a time of in- 
trusion and to a later metamorphic event. 
Other facets of the cooperative age pro- 
gram have included collaboration with 
Dr. Jager and Dr. Faul in measurements 
of Alpine rocks. 

Ore minerals, including those contain- 




ing Fe, Zn, Cu, Ni, Co, S, Se, and As, are 
under investigation by Kullerud, Arnold, 
Barnes, L. A. Clark, Roseboom, and Yund. 
Studies of the Fe-S, Cu-S, Ni-As, and 
Fe-As binary systems and Fe-S-O and 
Fe-Zn-S ternary systems were finished 
during the last three years, and the Ni-S, 
Fe-As-S, and Ni-As-S systems were com- 
pleted this year. The systems Fe-Se, Co-S, 
Fe-Ni-S, Cu-Fe-S, and Fe-S-Se are un- 
der continuing investigation. The results 
are furnishing an increasingly valuable 
means for interpretation of the relations of 
minerals in a large number of ore deposits. 

G. Donnay and J. D. H. Donnay have 
made interesting contributions to the 
theory of space groups as applied to struc- 
tures containing atoms carrying magnetic 
moments. This work is an extension of the 
theory of "colored space groups" developed 
by Belov in Russia, with whom the Don- 
nays have been collaborating. The exist- 
ence of polar magnetic vectors in a struc- 
ture introduces additional symmetry op- 
erations which increase the number of 
space groups from 230 to 1421. The deriva- 
tion of these new symmetry groups and 
their application to the ordering of mag- 
netic vectors in structures of the rock-salt 
type are described in more detail in a later 

The long-standing and petrologically 
important problems of relations between 
the pyroxenes, diopside, clinoenstatite, and 
pigeonite, as well as those between poly- 
morphs of pure MgSiOs — enstatite, pro- 
toenstatite, and clinoenstatite — have been 
studied in detail by Morimoto in coopera- 
tion with Appleman and Evans, of the 
U. S. Geological Survey. 

In his studies of mechanisms of heat and 
electrical conduction of the earth, S. P. 
Clark has shown that the strong ultraviolet 

absorption of iron-bearing olivines disap- 
pears when iron is not present in the 
mineral. This was an unexpected result, 
since previously the absorption was thought 
to be intrinsic to the silicate structure. 

Libby has studied radioactive fallout, 
particularly that associated with the Rus- 
sian high-latitude tests of October 1958. 
Using rainwater collected at the Geophysi- 
cal Laboratory and his facilities here he 
followed closely the time sequence of fall- 
out of Sr 89 and Sr no . The data are best 
interpreted on the basis of a residence time 
in the atmosphere of about 1 year. Thus, 
polar shots seem to lead to a time scale of 
fallout different from that of equatorial 
events, where the time is 5 years. 

Kerogen, by far the most abundant form 
of fossil organic chemicals, has been the 
object of two studies by Abelson. He has 
underlined the possible importance of re- 
actions between carbohydrates and pep- 
tides or amino acids in the formation of 
this complex constituent of sedimentary 
rocks. By low-temperature reduction with 
HI, kerogen has been split into petroleum- 
like substances including hydrocarbons. 

Measurable isotope fractionations are as- 
sociated with many important geological 
processes. A program of investigation in 
the field of isotope geology has been initi- 
ated at the Geophysical Laboratory. Hoer- 
ing, who joined our staff on January 1, 
1959, now has one mass spectrometer in 
successful operation and another well un- 
der way, together with the necessary aux- 
iliary gas trains and other equipment. The 
cooperation of Tuve, Aldrich, and Doak, 
of the Department of Terrestrial Magnet- 
ism, in the construction of the mass spec- 
trometer speeded Hoering's program about 
a year. 


Knowledge of the mineralogical constitu- 
tion of the upper mantle and of the deeper 
parts of the earth's crust is fundamental 
to progress in many fields of geophysics 
and petrology. Interpretation of seismic 

discontinuities and working out the origins 
of rocks such as eclogites, kimberlites, and 
granulites are only a few of the problems 
whose solutions depend in essential ways 
on our knowledge of mineral equilibria at 


great depth. Progress in this field depends 
directly on our ability to duplicate in the 
laboratory the conditions of high tempera- 
ture and pressure that exist at depth in the 
earth. Design developments outlined in 
last year's report now permit us to repro- 
duce the conditions of temperature and 
pressure equivalent to a depth of 150 kilo- 
meters with adequate accuracy and on a 
routine basis. We present below the re- 
sults of two mineral equilibria studies in 
this range. We have also made recon- 
naissance runs to pressures as high as 90 
kilobars, corresponding to a depth of about 
300 kilometers. 

Our most important study this year is a 
determination of the stability field of the 
magnesian garnet, pyrope. Much evidence 
suggests that pyrope-rich garnets may be 
major constituents of the rocks of the 
upper mantle. Our work on pyrope shows 
only one of a number of interesting effects 
of pressure on phase relations in the system 
MgO-Al 2 3 -Si0 2 . Others include the re- 
action cordierite ^ enstatite + sillimanite 
+ quartz which we have observed in the 
pressure range 10 to 20 kilobars, and the 
synthesis of enstatites containing over 10 
weight per cent AI2O3. Further study of 
this ternary system at high pressures will 
provide a better understanding of the con- 
ditions of formation of rocks belonging to 
the granulite facies. 

We are also presenting a study of the 
transition quartz ^ coesite. The inversion 
of quartz to coesite limits the depth in the 
earth at which quartz, a major constituent 
of the crust, can exist. This transition, 
which can be located within the precision 
of our apparatus over a temperature range 
of 1000° C, has been particularly useful in 
calibration studies. 

We have had to re-examine the problem 
of pressure measurement in the range up 
to 50 kilobars. This study has given us a 
much sounder understanding of the be- 
havior of solid pressure media at high 
temperatures and pressures, and we now 
believe that pressure can be measured in 
this range in our apparatus with an ac- 

curacy of ±5 per cent. This aspect of our 
work has led to a new determination of a 
transition in thallium at 37.1 kilobars of 
sufficient accuracy for use as a calibration 


F. R. Boyd and J. L. England 

Pyrope, Mg 3 Al 2 Si30i2, is the principal 
component in garnets from many eclogites, 
kimberlites, and granulites. The presence 
of pyrope-rich garnets in these rocks ap- 
pears to reveal the influence of high litho- 
static pressure during their formation. 
Granulites and many eclogites are clearly 
of crustal origin, but kimberlites and some 
eclogites evidently represent material 
brought from below the Mohorovicic dis- 
continuity. Pyrope-bearing eclogite and 
kimberlite may be important rock types in 
the upper mantle. Pure pyrope is not 
stable at atmospheric pressure. It can be 
synthesized at high temperatures and pres- 
sures, however, and study of its stability 
can advance our understanding of the 
origins of rocks in which pyrope-rich 
garnets have formed. 

A preliminary diagram of the stability 
field of pyrope is shown in figure 1. At 
pressures above 23 kilobars and tempera- 
tures above 1200° C, pyrope forms rapidly 
from glass or mixtures of anhydrous crys- 
talline phases. Most of our runs at 1200° C 
and all those below 1200° C were made 
with H 2 present as a flux. The H 2 
was allowed to escape during a run so that 
anhydrous products were obtained. Below 
1100° C reactions involving pyrope become 
sluggish, and curve A has not yet been 
extended to lower temperatures. In runs 
at pressures greater than 23 kilobars the 
yield of pyrope is over 95 per cent. The few 
per cent of other products persist as minute 
grains armored by garnet. The pyrope 
ordinarily forms grains with a diameter 
of about 0.01 to 0.05 mm; in one run at 
1300° C with H 2 present we obtained 
euhedral pyrope crystals with a diameter 
of 0.5 mm. The index of refraction and 
cell size of the pyrope synthesized in this 



investigation are given in table 1. They 
are in good agreement with values ex- 
trapolated from natural garnets and with 
values determined by Skinner (1956) 1 for 
pyropes synthesized by L. Coes and by 
E. C. Robertson. The presence or absence 
of H2O during a run made no difference 
in the refractive index of the pyrope. 

at atmospheric pressure. The stable as- 
semblage at atmospheric pressure is cor- 
dierite + forsterite + spinel. At pressures in- 
dicated for the runs to the left of curve A, 
cordierite is unstable and the stable as- 
semblage for the pyrope composition is 
aluminous enstatite + sapphirine [ + silli- 
manite ? ] . Enstatite forms 70 to 90 per cent 

a 1 . .-. DXX 

Aluminous enstatite / 

+ sapphirine n IS O ^ D 

(+ sillimanite ? ) / 


a 0/0 

Aa ° 


Pressure, kilobars 

Fig. 1. Stability field of pyrope, Mg 3 Al 2 Si30 12 . Shaded rectangular points represent glass or mix- 
tures of crystalline phases that reacted to pyrope. Open rectangular points with light outline are runs 
in which glass or mixtures of crystalline phases reacted to enstatite + sapphirine [+ sillimanite?]. 
X's are runs in which pyrope failed to break down; open rectangular points with heavy outline are 
runs in which pyrope broke down to enstatite + sapphirine [+ sillimanite?]. Circles are runs 
where pyrope formed from glass or mixtures of crystalline phases when seeded with 10 per cent 
crystalline pyrope. See text for explanation of field marked "Liquid (+ crystals?)." 

TABLE 1. Refractive Index and Unit Cell of 
Synthetic Pyrope 

n (Na) a Q , A 

Boyd and England 1.710±0.003 11.456±0.002 

Skinner (1956) 1.714 11.459 

Fleischer (1937)* ... 11.463 

Ford (1915)* 1.705 

* Values extrapolated from natural garnets. 

The phases stable for the pyrope com- 
position immediately to the left of curve A 
in figure 1 are not the same as those stable 

1 See References Cited at the end of this report. 

of the products of a run and hence appears 
to contain 10 to 20 weight per cent AI2O3. 
(It is hoped that X-ray studies will lead to 
a more accurate determination of the com- 
position of the enstatite.) 

The positions of the pertinent tie lines 
in the system MgO-ALOs-SiC^ have been 
fixed by runs on compositions other than 
pyrope at 15 to 20 kilobars. It is inferred 
from these studies that a small amount of 
sillimanite may be present with the ensta- 
tite and sapphirine on pyrope composition. 
We are in doubt about the presence of 
sillimanite because we have evidence that 



at these pressures and temperatures solid 
solution of sillimanite in sapphirine may 
be extensive enough to cause the pyrope 
composition to crystallize to the two-phase 
assemblage enstatite + sapphirine. 

We have encountered a remarkable nu- 
cleation effect in locating the position of 
curve A in figure 1. If glass or mixtures 
of crystalline phases are used as reactants, 
no pyrope is obtained in the region to the 
left of the dotted line marked D. At pres- 
sures higher than D, these reactants yield 
nearly 100 per cent pyrope. We initially 
believed that line D marked the boundary 
of the pyrope field, but further investiga- 
tion showed that pyrope could not be 
broken down until the pressure was re- 
duced below the range of curve A. A 
nucleation problem was suspected, because 
pyrope seemed to recrystallize in the field 
between curves A and D. Accordingly, a 
number of runs were made with glass 
seeded with 10 per cent crystalline pyrope. 
Pyrope grew readily from seed in these 
runs (circles in fig. 1), and hence curve A 
is the equilibrium boundary. 

Pyrope melts along curve B in figure 1. 
Melts from the temperature-pressure re- 
gion above curve B crystallize largely to 
enstatite + sapphirine [+ sillimanite?] in 
the quench. Little or no glass is obtained 
in the products. The identification of 
curve B as a melting curve rather than a 
subsolidus transition is based on its slope, 
the nature of the products quenched from 
above it, and the temperature range in 
which it is located. If B were a subsolidus 
transition, almost certainly products other 
than enstatite + sapphirine [ + sillimanite ? ] 
would have been obtained in the quench. 

The melting of pyrope must clearly be 
incongruent over an interval at the low- 
temperature end of curve B. In theory, B 
should not be a smooth curve, but should 
be inflected at junctions with invariant 
curves marking the successive disappear- 
ance of phases in the assemblage ensta- 
tite + sapphirine [ + sillimanite?] as the 
temperature is raised. Because of the im- 
possibility of quenching liquids in this 

system at high pressures and temperatures 
the sequence and positions of these curves 
cannot be established. Our data show only 
one invariant point at 1510° C and 21.6 
kilobars. Curve C, which intersects the 
pyrope stability field at this point, repre- 
sents the approximate position of one of 
the melting curves of the assemblage en- 
statite + sapphirine [ + sillimanite?]. Prob- 
ably C is a solidus curve and defines the 
melting of enstatite in this assemblage. Our 
data on the melting curve of pyrope (B) 
have been fitted by a straight line. It is 
probable, but not certain, that the melting 
of pyrope is congruent along most of the 
determined length of curve B. Runs 
quenched from above curve B contain uni- 
formly fine-grained quench products with- 
out large crystals that would be expected to 
form in suspension in liquid in the tem- 
perature range 1500° to 1750° C. 

The slope of the pyrope melting curve is 
14.1°/kilobar. This slope is similar to the 
slopes established for the melting of diop- 
side in the range 10 to 30 kilobars (10.3°/ 
kilobar; Boyd and England, Year Book 
57) and albite (ll°/kilobar; Birch and 
LeComte, 1958). The similarity in slope 
of these melting curves is surprising, since 
these minerals have very different com- 
positions and crystal structures. 

The position of the invariant point 
marked by the intersection of curves A, 
B, and C can be used to fix a maximum 
value for the rate of change with pressure 
of the minimum melting point in the sys- 
tem MgO-Al 2 03-Si0 2 . The temperature 
of the invariant point in figure 1 (1510° C) 
must be equal to or greater than the tem- 
perature of the minimum melting point in 
the ternary system at 21.6 kilobars. Since 
the temperature of the minimum in the 
ternary system at atmospheric pressure is 
1355° C (Schreyer and Schairer, this re- 
port), the maximum possible effect of pres- 
sure on the minimum melting point is 
about 7°/kilobar. This value is substan- 
tially less than the slopes of the melting 
curves of pyrope, diopside, and albite. Fur- 
ther data of this sort may show that the 


effect of pressure on the melting of poly- 
component rock systems will be substan- 
tially less than the 10°-14°/kilobar that 
would be predicted from present data on 
pure phases. 

The pyrope subsolidus boundary is 
shown in figure 2 with a variety of phase 

is intersected by reactions among the low- 
pressure products. Experimental and petro- 
graphic evidence suggests at least two as- 
semblages that might be stable along the 
pyrope boundary at lower pressures and 
temperatures than those investigated; these 
are forsterite + cordierite + spinel and en- 

Depth, kilometers ( p = 2.75) 

10 20 30 40 50 60 70 80 90 100 110 

120 130 140 

1000 - 


£> 800 — 

£ 600 


1 1 1 1 1 


1 1 1 1/ 1 


Enstatite + sapphirine / / 



(+ sillimanite?) 

\J / 



/ / 


/ / : 



4f / 

'/ / 

/ / 

/ Quartz ^coesite 

Albite + nepheline 4 

/ / 

5^ jadeite // , 


/ / 


/ // 


\ / 

Albite ^ 
jadeite + quartz 




1 i 1 . 1 

, I.I, 

200 - 

5 10 15 20 25 30 35 

Pressure, kilobars 

Fig. 2. Phase boundaries determined at high pressures and temperatures. These boundaries are 
extrapolated beyond their experimental ranges; see text for uncertainty in the position of the pyrope 
curve at low temperatures and pressures. Albite + nepheline ^ jadeite from Robertson, Birch, and 
MacDonald (1957); albite ^ jadeite + quartz from Birch and LeComte (1958); sillimanite ^ kya- 
nite from Clark, Robertson, and Birch (1957) and Griggs and Kennedy (1956). The quartz ^ coe- 
site transition is discussed in this report. 

boundaries determined in the pressure 
range above 10 kilobars. The quartz-coe- 
site transition is described in this report; 
the remaining curves have been taken from 
the data of other workers, and references 
to them are given in the legend of figure 2. 
Extrapolation of the pyrope boundary to 
temperatures below the range of experi- 
mental determination is subject to an un- 
certainty. It is possible that this boundary 

statite + cordierite + sapphirine. The effect 
of any reactions that might intersect the 
pyrope stability curve would be to bend 
the boundary toward higher pressures. The 
extrapolated boundary in figure 2 can be 
used as the best present approximation to 
the position of the pyrope stability curve 
at low pressures and temperatures, but it 
should be kept in mind that within ex- 
perimental error we can be certain only of 



the minimum pressure required to stabi- 
lize pure pyrope at any temperature below 
the experimental range. 

As is shown in figure 2, the pyrope sub- 
solidus boundary lies close to the curve for 
the reaction albite + nepheline ^ jadeite. 
Since eclogites consist essentially of a 
pyrope-rich garnet and a jadeitic pyroxene 
the pyrope and jadeite curves together can 
be taken as an approximation to the basalt- 
eclogite transition. Two uncertainties en- 
ter in this assumption. The effect of solid 
solution of almandine in the pyrope and 
diopside in the jadeite will be to lower the 
pressure necessary to form an eclogite. 
The addition of quartz to the system will 
have the opposite effect. Birch and Le- 
Comte (1958) have determined the influ- 
ence of quartz on the stability field of 
jadeite. Although further experimental 
work is needed to evaluate the other ef- 
fects quantitatively, some preliminary ap- 
plications can profitably be discussed. 

Robertson, Birch, and MacDonald 
(1957) have shown that the jadeite sta- 
bility curve and considerations of tem- 
perature versus depth in the earth's crust 
are consistent with the hypothesis that the 
Mohorovicic discontinuity is related to the 
change from the mineral assemblage of 
basalt or gabbro to that of an eclogite. If 
the pyrope ^enstatite + sapphirine [ + sil- 
limanite?] curve is extrapolated to low 
temperatures and pressures, the reaction to 
form pyrope is found to be in the tempera- 
ture range 300° to 350° C at a pressure 
equivalent to that at the base of the crust 
in a nonorogenic, continental area. These 
temperatures are at the lower end of the 
range calculated by Birch (1955) for vari- 
ous crustal models. The hypothesis of a 
basalt-eclogite transition at the Mohorovi- 
cic discontinuity seems worth continued 
consideration, although present data are 
too incomplete to provide a definitive 

Figure 2 shows that the pyrope and jade- 
ite stability boundaries are in the same 
range of pressures and temperatures as the 
kyanite-sillimanite transition. These data 

suggest that eclogite would be stable at 
any lithostatic pressure high enough to 
permit the formation of kyanite. Rocks of 
basaltic composition interbedded with 
kyanite schists nevertheless characteristi- 
cally crystallize as amphibolites rather than 
as eclogites. Ecologites of crustal origin 
must therefore either have formed at high 
temperatures above the stability field of 
hornblende at usual regional metamorphic 
vapor pressures or under rare conditions 
of particularly low vapor pressure at low 
temperatures. Most eclogites seem to have 
formed at relatively high temperatures, as 
judged from the degree of solid solution in 
feldspars and pyroxenes in associated rocks. 
Present experimental data certainly sug- 
gest that high lithostatic pressure is neces- 
sary for the formation of eclogite, but 
where eclogite is developed in regional 
metamorphic areas its formation in place 
of amphibolite is also related to high tem- 
perature or abnormally low vapor pres- 

In the course of our study of pyrope, we 
have noted the instability of cordierite at 
high pressures as well as the development 
of aluminous orthopyroxene. These are 
features characteristic of many granulites. 
Along with pyrope-rich garnets, these fea- 
tures are, perhaps, further evidence for the 
importance of high lithostatic pressure in 
controlling the development of mineral as- 
semblages in granulites. 

Quartz-Coesite Transition 
F. R. Boyd and f. L. England 

The dense polymorph of silica now 
known as coesite was one of the first sili- 
cate phases synthesized at high pressure 
and temperature which had not previously 
been discovered in natural occurrences or 
in the laboratory. Determination of the 
stability field of coesite is of geological in- 
terest because absence of coesite in crustal 
rocks makes it possible to set a limit on the 
depth of burial of the rocks in past ages. 
The quartz ^ coesite transition has also 
been useful because of the narrow pressure 


interval over which it can be reversed and 
the wide range of temperature over which 
the boundary can be located. It has been 
used repeatedly both to contrast results 
obtained with different pressure media and 
to test the consistency of results obtained 
with a pressure medium. 

A phase diagram showing the fields for 
all known stable phases of Si0 2 along 
with our runs on the quartz ^ coesite 
transition is presented in figure 3. These 


' Liquid ^'' 

i ■ I ■ 


J*"jt— Cristobolite 





-1 High Quartz 




1 Wridymite 

' : 1 



! 1000 

1 ,' odtt 
T / ' 7 •' 
■4 sill 

/ • J ■ 
4- / ' 9 ; 

r s 'll 

- / m 



- / 

' f I 


. / Low Quartz 


;' / / 

Ouorlz -*■ Coesite 
D Coesite-»-Quorlz 
D Ouortz, no chonge 
S Coesite; no change 










- ■ / 




• / .' 

, . 1 . 


1 . , . 1 . , ." 

10 20 30 40 50 

Pressure - Kilobars 
Fig. 3. The quartz ^ coesite transition. 

runs were made with talc as a pressure 
medium. The light dashed lines roughly 
paralleling the boundary represent limits 
of error of ±5 per cent of the pressure; 
these limits represent the maximum pres- 
sure uncertainty established by calibration 
on the transition Bi I ^ Bi II and Tl II ^ 
Tl III. The quartz ^ coesite boundary is 
linear over the temperature range 700° to 
1700° C and can be described by the equa- 
tion P=19.5 + 0.0112T, where P is in kilo- 
bars and T is in degrees centigrade. The 
reaction has been reversed within the pre- 
cision of our measurements over the en- 

tire temperature range investigated. Ex- 
trapolation of Yoder's (1950a) data on the 
high-low quartz inversion indicates a triple 
point high quartz-low quartz-coesite at 
approximately 1300° C and 34 kilobars. 
There should be a break in slope in the 
coesite stability boundary at this point, but 
our data are not precise enough to reveal it. 

It is possible that coesite will eventually 
be found in eclogites brought up from 
below the Mohorovicic discontinuity. The 
reaction of basalt to form eclogite can be 
modeled by the equation: 

Labradorite + olivine [+hypersthene] ^ 
pyrope-almandine + omphacite [ + Si0 2 ] 
If hypersthene is present in the reactants, 
free silica should appear in the products. 

We have checked these relations with 
several runs on basaltic glass from the 1952 
eruption of Halemaumau. This lava has 
a composition typical of the bulk of Hawai- 
ian basalts; its norm shows 1.5 per cent un- 
combined Si0 2 . At 1200° C and pressures 
ranging from 33 to 40 kilobars the basaltic 
glass crystallized to a mixture of garnet, 
clinopyroxene, minor calcic plagioclase, 
and 5 to 10 per cent free silica. The silica 
crystallized as quartz or coesite, depending 
on the pressure. Alderman (1936, p. 504) 
has described an eclogite from Glenelg 
which has a chemical composition nearly 
identical to the Halemaumau basalt. The 
norm of the eclogite shows 1.1 per cent 
uncombined silica, but the mode contains 
5.7 per cent quartz, in good agreement 
with our experimental results. 

Pacific basalts clearly originate at depth 
in the mantle. We do not know whether 
the mantle rocks from which basalts form 
have a chemical composition approaching 
that of the basalt or are more mafic. If 
they are not so mafic as to lack normative 
hypersthene, they may contain small 
amounts of modal coesite below a depth 
of about 110 kilometers. 

Thallium Transition 
F. R. Boyd and J. L. England 
Pressure calibration of our single-stage 
apparatus was initially based on measure- 


ments of the Bi I ^ Bi II transition at 24.9 
kilobars. This gave us a measurement of 
the friction at only one pressure, and it was 
desirable to obtain a second point at higher 
pressure. Bridgman's work (1935, 1952) on 
the transition Tl II ^ Tl III indicated that 
this transition lay in the right pressure 
range, and we accordingly undertook a de- 
termination of sufficient accuracy for use 
as a calibration point. 

We used the single-stage apparatus de- 
scribed in last year's report with a silver 
chloride cell shown in figure 4. A thallium 

, ,-, i ■ , . | i i i i | 


1 ' ' i-|-i-i ' ■ | 

0.013" thallium 
X — — — *^/ wire 


Gold lead ^- 1 

^-7^ ^AgCI 

Gold lead 



mperoture-30 °C 


1/4 inch 




Tl I[s=TI HI 
Bridgmon, 1935 

, , 1 . , ■ , 1 

• — Tl H=eTI ZK 
Bridgmon, 1952 
1 . . . . 1 



Pressure on piston ~ kilobars 

Fig. 4. Change in electrical resistance with 
pressure at the transition Tl II ^^ Tl III. 

wire 2 l / 2 inches long, 0.013 inch in di- 
ameter, was wound in a spiral between 
two disks of silver chloride and connected 
to insulated leads. The transition was de- 

tected by change in electrical resistance. 

A run with this type of cell is shown 
in figure 4. The abrupt change in resist- 
ance indicates the transition, and the dif- 
ference between the nick points on the 
curve with increasing pressure and with 
decreasing pressure measured along the 
pressure axis is the uncertainty interval 
within which the transition is located. By 
balancing on the transition with both Tl II 
and Tl III present, raising the pressure un- 
til II — > III then dropping the pressure until 
III— > II, the uncertainty interval can be re- 
duced below that shown in figure 4. 

We have made three setups for meas- 
uring this transition using two samples of 

TABLE 2. The Transition Tl II ^ Tl III, 
temperature 29° ± 1° C 


Pressure, Hysteresis, 

kilobars % 

N.B.S. (1) 
N. B. S. (2) 




thallium. One sample was obtained from 
the National Bureau of Standards and is 
99.9 per cent pure; the other was obtained 
from the Fisher Scientific Company and is 
listed as "purified." The results of these 
three runs with the hysteresis given as a 
percentage of the mean pressure at which 
the transition was observed with increas- 
ing and decreasing pressure are presented 
in table 2. The agreement is excellent, and 
the transition pressure at 29° C is 37.1 
kilobars ±3.5 per cent. 

H. S. Yoder, Jr., and C. E. Tilley 

Eclogite is a bimineralic rock consisting 
of an almandite-pyrope garnet and an 
omphacitic clinopyroxene. Because many 
eclogites are essentially equivalent in 
chemical composition to the various basalts 
and have a high density, they are con- 
sidered by some to be a high-pressure facies 


of basalt. The eclogites are therefore a 
potential primary source at great depths 
for basaltic magma. In fact, Fermor (1913) 
proposed an encircling layer of eclogite in 
the subcrustal zone of the earth from 
which small portions of eclogite are rarely 
transported intrusively or mechanically to 
the surface. Such a layer, if it exists, would 
be an adequate source for the large volume 


of basalt found on the surface of the earth. 
In order to study the potentiality of ec- 
logite as a source of basaltic magma, the 
melting behavior of two natural eclogites 
from Scotland was studied under anhy- 
drous and hydrous conditions at 10,000 
bars, the approximate pressure at the base 
of the continental crust. The composition 
of these eclogites is given in tables 3 and 4. 

data, 1958). The significance of these tie 
lines in outlining the limits of eclogite in 
relation to other mineral assemblages is 
given in the simplified projection of fig- 
ure 6. The diagram is a one-point projec- 
tion of SiO a onto the CaO-MgO-Al 2 3 
plane of the four-component system. 
Where the O'Hara assemblage fixes the 
limits of the eclogites in one respect, analy- 

TABLE 3. Analyses of Glenelg, Scotland, Eclogite (35090) and Its Garnet and Clinopyroxene, 
Norm of the Eclogite, and Atomic Ratios of Garnet and of Clinopyroxene 






Norm of 1 

Metal Atoms to 12(0) 

Metal Atoms to 6 (O) 

A1 2 3 






Si 2.9701 

Al 0.030J^ UUU 

Si 1-9161 

Al 0.084J AUUU 

Fe,0 3 






Al 1.9441 
Fe +3 0.056J 


Al 0.2701 
Fe+ 3 0.122 







Fe +3 0.046^ 

Fe+ 2 0.126 


Ti0 2 




H V 


Fe +2 1.476 

Mn 0.002 








Ti 0.019 
Ug 0.633 


Ti 0.018 
Mg 0.498, 

► 1.977 

Na a O 





Ca 0.759 
Mn 0.042, 

Ca 0.644 ) Q941 
Na 0.297 J u,y ^ 

H 2 + 




H O - 




p 2 o 5 


co 2 



26.i)3( A1 95 F e 5 ) 2 in atom 



per cent 




Less O 
= S 



1. Eclogite (35090), hills west of "Postman's Path" to Ardintoul, Glenelg, Scotland. 

2. Garnet from 1. 

3. Clinopyroxene from 1. 

All chemical analyses by J. H. Scoon, Cambridge University. 

The norm of the Loch Duich eclogite 
(Cambridge no. 35083) is equivalent to 
that of an iron-enriched olivine-rich tho- 
leiite, and that of the Glenelg eclogite 
(Cambridge no. 35090) is representative of 
a type of saturated tholeiite. The analyses 
of the coexisting garnet and clinopyroxene, 
which comprise more than 90 per cent of 
the rock, are also listed in the tables and 
are plotted in figure 5. The orientations of 
the garnet-clinopyroxene tie lines are com- 
pared, with kind permission, with those 
obtained by M. J. O'Hara (unpublished 

ses of garnet and clinopyroxene coexisting 
with kyanite fix them in another respect. 

The results of heating the Glenelg eclo- 
gite in the dry way at 10,000 bars indi- 
cated that clinopyroxene was the first phase 
to form on the liquidus at about 1225° C. 
(At 1 atm the eclogite has a clinopyroxene 
liquidus at about 1165° C; the liquidus 
rises, therefore, at the rate of about 6° C 
per 1000 bars.) Plagioclase appeared at ap- 
proximately 1175° C, and the liquid ap- 
pears to have completely crystallized at a 
temperature near 1025° C. The final prod- 


TABLE 4. Analyses of Loch Duich, Scotland, Eclogite (35083) and Its Garnet and Clinopyroxene, 
Norm of the Eclogite, and Atomic Ratios of Garnet and of Clinopyroxene 

Norm of 1 Metal Atoms to 12 (O) 


Metal Atoms to 6 (O) 

SiO z 






Al 2 O s 






Fe 2 3 


















Ti0 2 


















Na 2 





K 2 





H 2 + 




H 9 0- 

p 2 o 5 







Less O 




= S 







Al L9391 
Fe +3 0.061 J zm 
Fe +3 0.05 1" 
Fe +2 1.678 
Mn 0.042 
Ti 0.009 
Mg 0.645 
CaO 0.565 , 


Si L955 !-2 00 
Al 0.045 ] Im 


Fe +3 0.103 
Fe +2 0.230 
Mn 0.002 







0.7661 Q9V 
0.161/ u,yz ■ 



(Fe 57 .3Mn 1 .4Mg 22 . Ca 19 . 3 ) 3 (Al 95 Fe 6 ) 2 in atom 
per cent 

1. Eclogite (35083), % mile southeast of the Pier, Totaig, Loch Duich, Rossshire. 

2. Garnet from 1. 

3. Clinopyroxene from 1. 

All chemical analyses by J. H. Scoon, Cambridge University. 

AI 2 3 + Fe 2 3 -(No 2 + K 2 0) 


Mol per cent 

MgO + FeO + MnO + Ti0 2 

Fig. 5. Plot of analyzed coexisting clinopyroxene and garnet from eclogites of Glenelg, Scotland 
(35090), A and A' respectively; and of Loch Duich, Scotland (35083), D and D' respectively. These 
are compared with the limiting critical assemblage of clinopyroxene + garnet + orthopyroxene, 
H, H', and H" respectively; obtained with kind permission from M. J. O'Hara (unpublished data, 



ucts, clinopyroxene + plagioclase + magnet- 
ite, are the major phases of a basalt. In 
the temperature range 900° to 1250° C at 
10,000 bars, the eclogite behaved as a basalt; 
below 900° C the rate of reaction was too 
slow to determine whether or not eclogite 
was the stable assemblage. From above the 
liquidus the melt quenched almost com- 
pletely to a metastable clinopyroxene with 
some glass. This amazing behavior sug- 

the charge consisted entirely of clinopy- 
roxene! The analysis of this eclogite closely 
approximates a pyroxene formula (table 
6A). At 1100° C and below, the rate of 
reaction was insufficient to permit de- 
termination of the stable phases. 

Under hydrous conditions at 10,000 bars 
the eclogites were all liquid above 1025° C, 
and clinopyroxene and amphibole ap- 
peared together at approximately that tem- 




Mol per cent 

Hyp, 01 
FeO + MgO 

Fig. 6. Schematic diagram outlining limits of eclogite in relation to contiguous mineral assem- 
blages. Note that the solid solution of the garnets is limited by the anorthite-diopside assemblages. 
Gr, grossularite; An, anorthite; Wo, wollastonite; Aim, almandite; Py, pyrope; Di, diopside; Ol, oli- 
vine; Hyp, hypersthene; Ky, kyanite; and Sp, spinel. 

gests that pyroxenites of basaltic composi- 
tion may exist at depth. The analysis of 
the Glenelg eclogite can be reduced for the 
most part to a pyroxene formula (Table 

The melting behavior of the Loch Duich 
eclogite at 10,000 bars under anhydrous 
conditions exhibited unique characteristics. 
The liquidus was also 1225° C with clino- 
pyroxene appearing as the primary phase. 
(At 1 atm the Loch Duich eclogite liquidus 
is 1180° C with olivine.) Only 75° lower, 

perature. These phases were joined by 
sphene at a temperature near 700° C. Per- 
sistence of residual garnet and residual 
clinopyroxene of the eclogite prevented the 
attainment of equilibrium at lower tem- 
peratures in 24-hour runs. Above the liq- 
uidus the melt quenched entirely to an 
amphibole or an amphibole with some 
clinopyroxene. The reduction of the eclo- 
gite analyses, adding the requisite water, to 
an amphibole is presented in tables 5B and 
6B. The behavior of these eclogites is 


similar to that of basalt under hydrous con- 
ditions at 10,000 bars (see Year Book 55, 
p. 170). 

It is concluded from these exploratory 
studies that eclogite is not stable in the 

TABLE 5. Glenelg, Scotland, Eclogite Com- 
puted to Pyroxene and Amphibole Formulas 





0.004 [ 2.000 




0.926 J 







Fe +2 










►1.868 0.164 















A. Eclogite (35090), Glenelg, computed to py- 
roxene formula = 6(0). 

B. Eclogite (35090), Glenelg, computed to am- 
phibole formula = 24(0 + OH) after adding 
requisite water. 

earth's crust at temperatures above 700° C 
under either anhydrous or hydrous con- 
ditions. The unique hornblendites and 
highly aluminous pyroxenites obtained as 
quench products may now be included as 
potential source rocks for basaltic magma, 
and these are also apparently unstable in 
the same range of conditions. For the 
most part such rocks behave as basalt under 
similar conditions. At temperatures lower 
than 700° C, eclogite may become the sta- 
ble assemblage at pressures believed to 
exist in the crust provided that no water 
is present. That is to say, basalt may un- 
dergo a solid-state transformation to eclo- 
gite at some temperature below 700° C at 
10,000 bars under anhydrous conditions. If 
water is present eclogites are probably con- 
verted to amphibolites. At some pressure 
in excess of 10,000 bars (at depths of 60 to 
100 km) it is presumed that materials of 
basaltic composition crystallize as eclogite 

at all temperatures below the solidus under 
anhydrous conditions. 

On the basis of these conclusions eclogite 
is not a probable source of basaltic magma 
within the crust; at some depth below the 
crust, however, it is a potential source. The 
melting of eclogite must, of course, be 
complete to give rise to a basaltic magma. 
Partial melting of eclogite, the melting of 
some or all of the garnet or some or all of 
the pyroxene, will give rise, not to basaltic 
liquids, but to liquids which crystallize, 
for example, to a hypersthene allivalite or 
to nepheline pyroxenite, respectively (ta- 
ble 7) . Herein may lie the mechanism for 
producing at least two principal types of 
magma, one with tholeiitic characteristics 
and the other with alkaline affinities (see 
Tilley, 1958, pp. 336-337). The field asso- 
ciation of igneous eclogite with dunites 
and anorthosites suggests that melting or 
crystallization began in the subcrustal re- 

TABLE 6. Loch Duich, Scotland, Eclogite 
Computed to Pyroxene and Amphibole For- 




0.353 J 


1.705 J 




0.453 1 




Fe +2 







1 1.600 
r 1 - 956 0.432, 














A. Eclogite (35083), Loch Duich, computed to 
pyroxene formula = 6 (O). 

B. Eclogite (35083), Loch Duich, computed to 
amphibole formula = 24 (O + OH) after adding 
requisite water. 

gions and completed its course in the 
crust. The occurrence of eclogite nodules 
in some basalts (both oceanic and conti- 
nental) may be evidence of a similar his- 
tory. These observations are in agreement 



with the belief that some basaltic liquids 
originate at depths well below the Mohoro- 
vicic discontinuity. 

TABLE 7. Norms of Analyzed Garnets and 
Clinopyroxenes from Eclogites Given in 
Tables 5 and 6 

35090 35083 

















































E. G. Zies and F. Chayes 

Pseudoleucites — aggregates of feldspar 
and a sodic feldspathoid or zeolite with ex- 
ternal shapes suggesting that they are 
pseudomorphs of single crystals of leucite 
— have long been of interest as mineralogi- 
cal curiosities. Although descriptions and 
chemical analyses of them may be found 
in the scientific press of more than a 
century ago, even petrographers continued 
to regard them primarily as oddities until 
Bowen's ingenious speculations gave them 
a central role in arguments about the 
evolution of alkaline rocks and magmas. 

Bowen (1928, 1937) contended that 
pseudoleucites were the natural record of 
incongruent melting relations known to 
exist for pure KAlSiyOs and encountered 
in laboratory studies of the system NaAl- 
Si04-KAlSi04-Si02, and that an under- 
standing of the consequences of this in- 
congruence was vital not only to proper 
evaluation of phase equilibria and frac- 
tionation relations in the synthetic system 

but also to an adequate theory of the 
origin of alkaline rocks and magmas. It is 
probably safe to say that he was the only 
petrologist who ever wholly believed his 
pseudoleucite hypothesis, and that since 
its announcement few serious students of 
the alkaline rocks have wholly disbelieved 

Under the circumstances, thorough mi- 
croscopic and chemical analysis of pseudo- 
leucites and the rocks in which they occur 
is of considerable importance to penolo- 
gists. We have for some time been in- 
volved in such an analytical study of a 
particularly fine specimen from the Bear- 
paw Mountains, Montana. During the past 
year we have revised our sampling pro- 
cedure, repeated on new material most of 
the work reported last year, and extended 
the sampling to include the groundmass of 
the rock. 

The new sampling was made primarily 
to test our proposed explanation of the 
discrepancy between micrometric and 
chemical estimates of the mode (see Year 
Book 57, pp. 204-206). Interested readers 
may recall that, although nepheline is the 
only acid-soluble constituent of any conse- 
quence, the weight loss on acid extraction 
was only about two-thirds of the nepheline 
content estimated by modal analysis made 
in thin section. We attributed this dis- 
crepancy to loss of nepheline in the sizing 
which preceded magnetic separation of 
pseudoleucite from groundmass. A frag- 
ment count on the actual powder used for 
chemical analysis supported this sugges- 
tion, but very little of the powder could be 
spared and its wide range of grain size 
made it far from optimum for this purpose. 
The question demanded further attention, 
especially since we had originally chosen to 
work on this particular specimen because it 
seemed to offer an excellent opportunity for 
cross checking of micrometric and analyti- 
cal procedures. 

A new sampling was accordingly car- 
ried through in duplicate, entirely by hand 
sorting under a magnifying glass or low- 


power microscope; acid extraction data for 
the two new products are shown in table 8. 
The average nepheline estimated micro- 
metrically from eleven thin sections (Year 
Book 57, p. 205, table 6) is 29.8 per cent. 
The results are thus in reasonable agree- 
ment, and we feel confident that our ex- 
planation of the very much larger dis- 
crepancy reported last year is essentially 
correct. Although the reproducibility of 
chemical results on the final powder is 
certainly far superior to micrometric pre- 
cision, columns 1 and 2 of table 8 are a 
reminder that chemical procedures are sub- 
ject to the same sampling variation en- 
countered in micrometric work. The point 
is almost self-evident and would scarcely 
be worth mentioning were it not so con- 
sistently ignored in the journal literature. 

TABLE 8. Extraction of Two Samples of Pseu- 
doleucite in 1 : 5 HC1 


2 Average 

Weight loss to acid, per 

cent of whole 29.0 27.4 28.2 

Weight recovered from 

acid, per cent of whole 29.2 27.3 28.2 

Since a full report of this work will soon 
be submitted for publication, we present 
here only a summary of our principal find- 

Nepheline. The nephelines of ground- 
mass and "pseudoleucite" are virtually 
identical. They contain a little more than 
15 per cent Na 2 and a little less than 8 
per cent K a O. In neither is there a sig- 
nificant departure from the 1:1:2 ratio of 
alkalies, R2O3, and silica, and both contain 
a little over 2 per cent Fe 2 3 . 

Feldspar. The feldspars of groundmass 
and "pseudoleucite" are also extraordinar- 
ily similar. Both are exceedingly potassic 
sanidines (Or 97 in pseudoleucite feldspar, 
Or 94 in groundmass feldspar). Both are 
very poor in soda. The biggest difference 
between them is in BaO content, which 
calculates to about 1 per cent of celsian 
molecule in the pseudoleucite sanidine as 

compared with 3 per cent in the ground- 
mass sanidine. 

Modal ratios of nepheline and feldspar. 
In the system Kp-Ne-Si0 2 our Bearpaw 
pseudoleucite composition of course pro- 
jects on a line connecting the plotted values 
of the nepheline and feldspar analyses. Both 
chemical and micrometric modes place it 
very nearly at the intersection of this line 
with a line joining pure Lc and the Bowen 
pseudoleucite reaction point. (It is also, of 
course, very close to the join between Lc 
and Na 2 0-Al 2 3 -2Si0 2 , the projected 
composition of pure Na-analcime.) 

Whether the initial phenocrysts are leu- 
cite or analcime, if a reaction of the type 
envisaged in Bowen's pseudoleucite hy- 
pothesis proceeds to completion, phases 
common to pseudomorphs and ground- 
mass will have the same composition in 
each; as far as major constituents are con- 
cerned this is true of nepheline and feld- 
spar in our specimen. 

With regard to the compositions of 
nepheline and feldspar to be expected, 
however, the hypothesis is not very specific. 
This is partly because, at the time it was 
proposed, information about phase rela- 
tions in "petrogeny's residua system" was 
insufficient to permit detailed prediction. 
Extensive current researches are gradu- 
ally eliminating this difficulty, but it seems 
to us that the very restriction of the argu- 
ment to iron-free systems is a considerable 

Our pseudoleucite, like nearly all those 
for which adequate descriptions are avail- 
able, occurs in an iron-rich rock which con- 
tains little magnesia. It seems quite pos- 
sible that the oxidation state of iron at the 
time leucite becomes unstable exerts a de- 
cisive influence on the further history of 
the rock. If the iron is ferrous and mica 
can form, as in a number of plutonic as- 
semblages, both alkalies may be fixed in 
dark silicates, potash predominantly in 
mica and soda in amphibole, so that an 
albite-microcline-nepheline or even an al- 
bite-nepheline assemblage may result. If, 
as in many volcanic assemblages, iron is 



dominantly ferric and only pyroxene and 
amphibole are stable, much soda but very 
little potash will be fixed in the dark sili- 
cates. The potash must be accommodated 
somewhere, with the rather paradoxical re- 
sult that an extremely potassic feldspar may 
precipitate from a residual liquid rich in 
soda. The final assemblage may thus be 
what we find in our specimen, viz., acmite, 
nepheline, and sanidine, the first virtually 
free of K, the second carrying as much K 
as the temperature will permit, the third, 
the most abundant mineral of the rock, a 
virtually pure K end member of an Na-K 
solid solution series. 


Cordierite occurs in igneous as well as 
metamorphic rocks of proper bulk compo- 
sition. Its presence or absence in these 
rocks may yield valuable information 
about the physical conditions the rocks 
endured before exposure at the earth's sur- 
face. The physical conditions, of course, 
can only be deduced if the maximum sta- 
bility range of the mineral is known quan- 
titatively. For this reason extensive labora- 
tory experiments on pure Mg cordierites 
under varying pressures and temperatures 
were carried out. In order to correlate the 
experimental results with natural occur- 
rences, a review of the geological and 
petrological literature on cordierite-bearing 
rocks was undertaken in connection with 
field studies and microscopic examinations 
of natural rocks. The sections to follow 
record the progress made in: (1) field and 
microscope studies of natural cordierite- 
bearing rocks, (2) heating experiments on 
and near the anhydrous Mg-cordierite 
composition in the MgO-Al 2 03-Si0 2 sys- 
tem, and (3) hydrothermal studies on the 
Mg-cordierite composition. Additional 
studies on the nature of cordierite which 
coexists with garnets may be found in the 
section on garnets. 

Natural Cordierite-Bearing Roc\s 
W. Schreyer 
During the summer of 1958 a number 
of cordierite-bearing rocks were collected 

from several localities in the United States. 
The mineral assemblages encountered and 
the main problems involved are briefly 
outlined below. 

Near Guilford, Connecticut, cordierite 
occurs as round, mostly noneuhedral crys- 
tals up to 2 inches in diameter in peg- 
matitic quartz-plagioclase veins largely 
following the schistosity of a very biotite- 
rich gneiss which itself contains minor 
amounts of cordierite. 

In the area south of Central City, Colo- 
rado, a quartz-monzonite gneiss contains 
lenses of rocks with peculiar bulk composi- 
tions. Some of them consist only of cordi- 
erite and anthophyllite or gedrite and a 
little quartz; others, of cordierite, garnet, 
and anthophyllite with a little quartz. 
Seemingly unlikely assemblages like cor- 
dierite + anthophyllite + garnet + spinel + 
quartz or cordierite + corundum were 
found. Metasomatic processes appear to be 
necessary to account for such curious non- 
equilibrium assemblages. These rocks 
have strong similarities to those from 
Orijarvi, Finland, and Kongsberg, Nor- 
way (Bugge, 1943). 

Cordierite-bearing rocks from the Lara- 
mie Range, Wyoming (Newhouse and 
Hagner, 1949), are characterized primarily 
by the assemblage cordierite-hypersthene- 
spinel, which seems to indicate the absence 
of water and alkalies. By the introduction 
of alkalies, water, and more silica at a 
later stage of the metamorphism this as- 
semblage was gradually replaced by bio- 
tite-plagioclase-K feldspar-quartz rocks of 
more and more granitoid composition. Re- 
sidual cordierites in these rocks have a core 
of green spinel and are surrounded by a 
bright quartz-plagioclase-K feldspar rim 
which is poor in biotite. Again, metaso- 
matic changes of the bulk composition of 
the rocks must be presupposed. 

East of Texas Creek, Colorado (Travis, 
1956), as well as just east of the Harding 
Mine near Dixon, New Mexico (Mont- 
gomery, 1953), cordierites occur as very 
large prismatic crystals of 6 inches and 
more in quartz-muscovite schists, with 


minor amounts of plagioclase, garnet, bio- 
tite, and andalusite. Near Texas Creek 
large sillimanites are also found. The cor- 
dierites are porphyroblasts, which contain 
about 30 per cent quartz and ore minerals 
as inclusions of the matrix of the country 
rock. Muscovite and biotite are rarely in- 
cluded within the cordierites. They seem 
to have participated in the reaction to form 
cordierite. The cordierites of these two lo- 
calities, especially the ones in the New 
Mexico rock, form round cigar-shaped 
crystals, elongated parallel to their c axis, 
which are oriented linearly in the rock 
parallel to a fold axis (B) . The inclusions 
in the cordierites are arranged in two 
planes (internal S planes), which intersect 
in the c axis of the crystal or B axis of the 
rock respectively. Nearly always these in- 
ternal S planes form rather large angles 
with the external S plane of the quartz- 
muscovite schists, yielding structures simi- 
lar to those of rolled garnets. If the struc- 
tural interpretation is right, the cordierites 
were stable or even actually growing in a 
stress field which resulted in a rotation of 
the crystals around their c axes, a relation 
hardly compatible with the hypothesis that 
cordierite is strictly an antistress mineral. 

In the area around Gillette, Wyoming, 
burning coal fields of Tertiary age have 
under surface conditions produced a series 
of pyrometamorphic rocks, locally named 
clinkers, within the adjoining shaly coun- 
try rocks. The clinkers are extremely fine- 
grained, gray or reddish rocks with a 
slaglike appearance. A number of them 
show flow structures, indicating that they 
crystallized from a melt. Fermor (1924) 
has coined the name paralava for similar 
rocks from the Bokaro Coalfield, India. 
The mineral assemblages in the Wyoming 
rocks as determined by X-ray diffraction 
patterns are made up mainly of tridymite, 
cordierite, and plagioclase, with rare mull- 
ite and quartz. Miyashiro and Iiyama 
(1954) first discovered a hexagonal poly- 
morph of cordierite in the Indian rocks 
and named it indialite. The cordierites of 
the Wyoming clinkers displayed a series 

of intermediate structural stages between 
this hexagonal and the usual orthorhombic 
form. True hexagonal indialites have not 
been found in these rocks as yet. 

Of particular interest are the cordierite- 
containing emery deposits of the Cortlandt 
complex, New York (Williams, 1888; 
Bowen, 1922; Friedman, 1956; and many 
others) . The assemblage cordierite-spinel, 
common in most cordierite-bearing rocks, 
is not found here. The join is broken 
by the join garnet-sillimanite and/or by 
the presence of the mineral sapphirine, 
which seems to have the composition 
2(Mg,Fe)02Al 2 CvSi0 2 (Friedman, 

1954) . Cordierite-garnet-sillimanite-sap- 
phirine assemblages are found in gray 
emery, but occur also as reaction rims 
around quartz masses in black emery, 
which is made up of magnetite, spinel, and 
corundum. According to the new data on 
the stability of cordierite Friedman's 
(1954) conclusion that the cordierite-bear- 
ing assemblages are due to a secondary 
reaction of the black emery with later 
quartz veins does not appear to be valid. 
The principal objection is that the veins 
of pure quartz are only formed under 
much lower P-T conditions, most probably 
as precipitates of silica-bearing hydrous 
solutions. It is suggested, therefore, that 
these quartz masses, which form pods, 
lenses, and intensely folded veinlets actu- 
ally included in the emery (compare with 
Friedman, 1956, fig. 10; 1954, fig. 1), were 
present before the formation of the emery. 
In concordance with Bowen's (1922) ex- 
planation of the emery as desilicified and 
dealkalized relics of Manhattan schists in 
the norite, it is suggested that the border- 
ing sapphirine-sillimanite-garnet-cordier- 
ite rock was formed under the high tem- 
peratures of contact metamorphism by 
successive reaction of the completely desi- 
licified black emery with the highly refrac- 
tory original quartz veins of the Manhat- 
tan schist. 

The help of Robert H. Moench, Frank 
W. Osterwald, Paul K. Sims, Ogden L. 
Tweto, all of the U. S. Geological Survey, 


Denver, Colorado; Fred Barker, U. S. Geo- 
logical Survey, Washington, D. C; Arthur 
Montgomery, Lafayette College, Easton, 
Pennsylvania; and J. Frank Schairer of this 
Laboratory, in locating and collecting the 
above-described rocks, is greatly appreci- 

Cordierites in nature are very often al- 
tered into greenish products usually called 
pinite. X-ray studies on a number of speci- 
mens (some of which were kindly made 
available by Professor G. Fischer, Munich, 
and Dr. W. Wimmenauer, Freiburg i. Br., 
Germany) showed that they usually con- 
sist of an intimate intergrowth of a mus- 
covite and a chlorite. Early stages of the 
pinitization exhibit a 7 A phase which may 
be called aluminous serpentine after Yoder 
(1952), or septechlorite after Nelson and 
Roy (1958). Late stages of pinitization 
contain normal 14 A chlorites. Neither 
pyrophyllite nor montmorillonite has as 
yet been found. Apparently the break- 
down of cordierite in nature usually in- 
volves K 2 0, and an investigation of the 
stability relations of cordierite in the pres- 
ence of K 2 is desirable. 

Anhydrous Cordierites and the System 
MgO-Al>O s -SiOo 

W. Schreyer and J. F. Schairer 

The melting and subsolidus relations of 
43 compositions in the system MgO- 
AL0 3 -Si0 2 have been investigated with 
the intention of determining possible solid 
solution and polymorphic changes in the 
ternary compound cordierite. 

Composition. Glass of the composition 
2Mg02Al 2 Gv5Si0 2 (2:2:5) crystallizes 
completely to cordierite at subsolidus tem- 
peratures. Upon crystallization, composi- 
tions containing 2.6 per cent spinel, or 2 
per cent forsterite, or 6 per cent pyroxene 
(the polymorphic form of MgSi0 3 will be 
disregarded) in addition to 2:2:5 all 
showed the appropriate amounts of these 
compounds along with the cordierite. It 
can be stated, therefore, that crystalline 
material of the composition 2:2:5 shows 

no variation toward MgSiO*3, forsterite, or 

In studies to ascertain the possibility of 
solid solution of cordierite toward silica 
(Rankin and Merwin, 1918), difficulties 
were encountered because of the sluggish- 
ness of the crystallization of silica. Trid- 
ymite, the stable phase of silica at sub- 
solidus temperatures above 867° C, never 
formed from any of our compositions lying 
along the line 2:2: 5— silica. Omitting the 
metastable formation of a high-quartz 
structure as previously reported (Year 
Book 57, p. 198), the first pure silica phase 
crystallizing was always found to be the 
high-temperature form of cristobalite, 
which has its stable range between 1470° 
and 1710° C, well above our crystallization 
temperatures. From silica-rich composi- 
tions various amounts of the metastable 
low-temperature form of cristobalite were 
obtained also. 

Silica-rich compositions on the line 
2:2: 5— silica readily crystallize to a mixture 
of cristobalite and indialite (hexagonal 
form of cordierite) with metastable prod- 
ucts present also, if the crystallization times 
are short, for example a few days at 1250° 
C. Compositions close to 2:2:5, such as 
1:1:3, on the other hand, seem to crystal- 
lize completely to indialite under these 
same conditions. No residual glass can be 
detected optically. But after crystallization 
of 2 weeks or more at temperatures be- 
tween 1350° and 1400° C cristobalite can 
be seen optically as tiny pinpoints in the 
cordierite, and X-ray patterns show a peak 
of high-cristobalite at 35.7° 20. It is con- 
cluded, therefore, that there is no solid solu- 
tion in cordierite (2:2:5) toward silica. It 
is likely, however, that the indialite formed 
from the glass at low temperatures contains 
excess silica in metastable solid solution. 
At temperatures between 1400° C and the 
beginning of melting of the 2:2: 5— silica 
compositions at about 1440° C a gradual 
decrease in the amount of cristobalite can 
be observed. This may be due to prema- 
ture formation of small amounts of liquid 



because of impurities (alkalies) in the 
chemicals. There is no substantial evi- 
dence that the partial disappearance of 
silica in this temperature range is caused 
by solid solution. 

It was stated by Iiyama (1955), on the 
basis of experimental work, that an al- 
most complete series of solid solutions 
exists between indialite 2:2:5 and a theo- 
retical compound "Mg beryl," Mg 3 Al 2 Sie- 
Ois (3:1:6). Our own data on composi- 
tions along this line show clearly that the 
solid solutions observed by Iiyama have to 
be considered metastable, owing to short 
periods of crystallization (10 hr, 60 hr). 
Furthermore, Iiyama's crystallization tem- 
perature of 1350° ±20° C for four of his 
six compositions was inadequate, because 
the equilibrium assemblage for these com- 
positions, except for 2:2:5 itself, above 
1355° ±3° C (see below) is cordierite in 
glass without any pyroxene or silica. All 
our compositions, which included one very 
close to 2 : 2 : 5 (93.2 per cent 2 : 2 : 5, 4.1 per 
cent MgSiOs, 2.7 per cent Si0 2 ), when 
crystallized for long periods below 1350° C 
consisted of the three phases cordierite, 
pyroxene, and a silica phase. 

Quenching runs on a number of compo- 
sitions within the cordierite field indicated 
that their paths of crystallization were, 
within experimental error, straight lines 
leading away from 2:2:5 composition. 

We conclude, therefore, that the compo- 
sition of cordierite in the system MgO- 
AI2O3-S1O2 is 2Mg02Al 2 CV5Si0 2 and 
that there is no solid solution in any direc- 
tion within experimental error. 

Invariant points. In the course of our 
investigations new data were obtained de- 
fining more precisely the temperatures and 
compositions of some of the invariant 
points in the system MgO-Al 2 Os-Si0 2 . 
Schairer (1954), on the basis of data ob- 
tained in the system K 2 0-MgO-Al 2 3 - 
Si0 2 , has pointed out that the temperature 
1425° ±5° C for the ternary point tridy- 
mite-mullite-cordierite in the system 
MgO-Al 2 0.i-Si0 2 as given by Rankin and 
Merwin (1918) must be too low. He con- 

cluded that it should lie between 1435° ±5° 
and 1448° ±5° C. In our present study 
quenching runs on six compositions lying 
close to this ternary point yielded data that 
indicate that its temperature is 1440° ±3° 
C and its composition 9.25 MgO, 22.50 
AI0O3, 68.25 Si0 2 . 

Quenching runs on four compositions 
lying near the ternary eutectic protoen- 
statite-tridymite-cordierite yielded data 
that indicate a temperature of 1355° ±3° C 
and a composition 20.50 MgO, 17.50 AI2O3, 
62.00 Si0 2 for this invariant point. 

Eleven compositions along the line cor- 
dierite-MgSiOs were studied, and the tem- 
perature at which cordierite is joined by 
protoenstatite, or protoenstatite is joined by 
cordierite, was determined as 1365° ±2° C. 
Since the position of this point is at the 
intersection of the join cordierite-MgSiO:; 
with the boundary curve between the cor- 
dierite and protoenstatite field it has to be 
a temperature maximum on this boundary 
curve. This is 10° above the temperature 
of the ternary eutectic protoenstatite-cor- 
dierite-tridymite given above and 5° above 
that of the ternary point protoenstatite- 
forsterite-cordierite given as 1360° ±5° C 
by Rankin and Merwin (1918) ; the ternary 
point protoenstatite-forsterite-cordierite, 
therefore, must be a ternary eutectic. 

The temperature of the maximum on 
the boundary curve cordierite-mullite, 
which is at the intersection of the exten- 
sion of the join mullite-cordierite with 
this boundary curve, was determined as 
1465° ±2° C. It lies between the ternary 
reaction point tridymite-mullite-cordierite 
(1440° ±3° C as given above) and the 
ternary reaction point sapphirine-mullite- 
cordierite (1460° ±5° C as determined by 
Keith and Schairer, 1952). The slope of 
the boundary curve from this maximum in 
both directions was found to be very 
gentle in the vicinity of the maximum 
itself, becoming steeper toward the two 
ternary points. 

Polymorphism. A particular effort has 
been made to obtain data on the relations 
between orthorhombic pseudohexagonal 



cordierite (two- or multipeak form) and 
hexagonal indialite (single-peak form) as 
first distinguished by Miyashiro and 
Iiyama (1954). It can be said with cer- 
tainty that the indialites made from glasses 
of all our compositions at low temperatures 
exhibit metastable stages in the crystalliza- 
tion history. Given sufficient time, all 
these indialites will eventually invert to 
multipeak cordierites, going through a 
complete series of intermediate structural 
stages. This gradual structural inversion 
is the more sluggish the farther away the 
bulk composition lies from that of pure 
cordierite 2:2:5. In all compositions in- 
vestigated except in those on the line 
2:2:5-spinel the highest rate in producing 
multipeak cordierite was obtained with 
temperatures between 20° and 40° below 
the temperatures at which the particular 
bulk composition begins to melt. Never- 
theless heating periods of almost 2 months 
were required in several cases. It is con- 
cluded that for these compositions multi- 
peak cordierite is the stable polymorph at 
least up to the beginning of melting. 

Pure cordierite 2:2:5 theoretically melts 
to mullite 4- liquid at 1465° C. In practice 
we always observed the development of 
mullite and liquid as low as 1440° C, 
which increased in amount until cordierite 
disappeared at 1465° C. Multipeak cordier- 
ite used as starting material in runs of 24 
hours to 7 days went through a gradual 
structural inversion to one-peak indialite in 
the temperature range between 1450° and 
1460° C. For the 1:1:3 composition this 
inversion was found to lie between 1440° 
and 1450° C. The gradual disappearance 
of cristobalite in this composition below 
1440° C, therefore, cannot be related to the 
structural change of cordierite back to 

The same type of structural inversion is 
also found in most of the other composi- 
tions investigated. The single-peak form 
is encountered at temperatures slightly be- 
low that at which cordierite disappears, 
which theoretically may be as low as 
1355° C, the temperature of the ternary 

eutectic cordierite-pyroxene-silica. In all 
these cases the formation of much liquid 
seems to be the determining factor for the 
inversion. Compositions close to ternary 
points or those in the pyroxene or forsterite 
or silica fields, which develop liquid theo- 
retically at the same temperature at which 
they lose the phase cordierite, should there- 
fore show the inversion over an extremely 
narrow range of temperature. In practice 
we found that these compositions usually 
show the multipeak form up to the com- 
plete disappearance of the phase cordierite 
despite the presence of much liquid. 

An exception to the observation that the 
inversion takes place only in the presence 
of liquid is exhibited by compositions 
along the line 2:2: 5-spinel. Four of these 
were available for study with 2.6, 6.5, 8.5, 
and 19.6 per cent spinel. These composi- 
tions do not begin to melt until 1453° ±5° 
C, the temperature of the ternary reaction 
point cordierite-sapphirine-spinel. The 
gradual inversions, however, take place for 
the 2.6 per cent spinel mixture around 
1400° C, for the one with 6.5 per cent 
around 1350° C, with 8.5 per cent around 
1300° C, and with 19.6 per cent below 
1050° C. One is tempted to believe, there- 
fore, that increasing amounts of spinel 
stabilize the single-peak form. 

These data indicate that the temperature 
of the cordierite-indialite transition is 
greatly influenced by the bulk composition 
of the particular mixture. In view of the 
complexity of the bulk compositions of 
natural cordierite-bearing rocks, the use of 
this transition as a geological thermometer 
is considered very impractical. 

Cordierite-W ater System 
W. Schrcyer and H. S. Yoder, Jr. 

The stability limits of the magnesium 
end member of cordierite have been deter- 
mined at 1000, 2000, 5000, and 10,000 bars 
water pressure. A preliminary P-T dia- 
gram is given in figure 7. 

Lower stability limits. The lower break- 
down products of Mg cordierite at 2000 



and 5000 bars are pyrophyllite and amesite 
or montmorillonite + a chlorite or only a 
montmorillonite solid solution at succes- 
sively lower temperatures (Year Book 57, 
p. 196). The reaction rates in the vicinity 
of the alleged breakdown curve are ex- 
tremely sluggish. Synthetic cordierite held 
at 450° and 500° C at 2000 bars and 525° C 
at 5000 bars for 4 to 8 months showed no 
signs of breakdown, and cordierite was not 

center of this range of indifference. Its 
curvature below 2000 bars was calculated 
using the values at 2000 and 5000 bars. The 
equilibrium value at 1 bar Ph 2 o was found 
to be 310° C. Although the formation of 
cordierite was never complete below 600° C 
at 2000 bars and 650° C at 5000 bars the 
by-products did not include pyrophyllite. 
The aluminosilicates were represented by 
andalusite or andalusite + mullite in addi- 

t— i re v »r 

Chlorile + Quartz + 
"Sillimanite" + Vapor 

Montmorillonite + 
Chlorite + Vapor 

Solid Solution 
+ Vopor 

Cordierite + Vapor 

200 400 600 800 1000 

Temperature, °C 

Fig. 7. Preliminary P-T diagram of the system Mg cordierite-water. Crosses indicate growth of 
cordierite; crosses in boxes, breakdown into pyrophyllite + amesite; open boxes, neither growth nor 
breakdown of cordierite; open triangles, breakdown into chlorite + montmorillonite or montmoril- 
lonite solid solution only; solid dots, breakdown into chlorite + quartz -f- metastable corundum; 
crosses in circles, incongruent melting products -(-liquid ± cordierite; open circles, liquid only. 
Minor changes in slope of the beginning of melting curve caused by the various incongruent melt- 
ing products have not been indicated. 

formed from the most reactive starting ma- 
terial (glass) under the same conditions. 
This indicates a range of indifference of at 
least 50° C at 2000 bars which becomes nar- 
rower with increasing water pressure. Mg 
cordierite glass held at 550° C and 2000 
bars for 2 months yielded only about 60 
per cent cordierite in addition to other 
products. Crystalline cordierite held at 
400° C and 2000 bars for 3 months showed 
very slight alteration to montmorillonite. 
On the basis of these observations the P-T 
curve, therefore, was drawn through the 

tion to quartz and frequently a little co- 
rundum. This shows that equilibrium was 
not attained even in runs of several months' 
duration. Long runs usually yielded much 
more andalusite than mullite. It is con- 
cluded, therefore, that andalusite is the 
stable aluminosilicate under these P-T con- 
ditions. Nevertheless the formation of an- 
dalusite from the cordierite bulk composi- 
tion has to be considered metastable be- 
cause of the slow growth of cordierite. All 
the runs on cordierite glass above 600° C 
at 2000 bars and 650° C at 5000 bars yielded 



all cordierite except for small amounts of 
leach products. 

Different products are encountered at 
10,000 bars water pressure. Here cordierite 
breaks down readily at about 685° C, well 
above the stability of pyrophyllite (about 
575° C for a cordierite bulk composition). 
In the range from 600° to 675° C the as- 
semblage chlorite + quartz + corundum was 
obtained after 4 days from the breakdown 
of crystalline cordierite as well as from 
cordierite glass. Natural pyrophyllite run 
at 675° C, 72 hours, broke down into 
quartz + corundum as well. Natural kya- 
nite and sillimanite under the same condi- 
tions showed no sign of alteration; syn- 
thetic mullite, however, broke down into 
sillimanite + corundum. It seems likely, 
therefore, that the stable aluminum silicate 
in this field is sillimanite. The position of 
the invariant point, at which cordierite is 
in equilibrium with chlorite, pyrophyllite, 
sillimanite, quartz, and vapor, has not been 
fixed. Its location will help in determining 
the slope of the univariant curve chlorite + 
sillimanite + quartz ^ cordierite + vapor. 
The intersection of this curve with the 
melting curve of cordierite (described be- 
low and in fig. 7) will determine the termi- 
nation of the cordierite field toward higher 
water pressures. Data obtained in the sta- 
bility range of cordierite at 10,000 bars 
water pressure suggest that this univariant 
curve is rather flat. The rates of formation 
of cordierite at these high water pressures 
are amazingly slow. Cordierite glass held 
at 800° C, 10,000 bars, for 24 hours yielded 
only talc + sillimanite + quartz; after 60 
hours the charge consisted of cordierite, 
talc, quartz, and corundum. This may in- 
dicate the proximity of a reaction curve, 
and therefore cordierite may not be stable 
at water pressures slightly higher than 
10,000 bars. Preliminary runs on water-free 
cordierite composition at confining pres- 
sures from 13 to 20 kilobars in the Boyd 
and England apparatus resulted in the 
breakdown of cordierite and formation 
of assemblages which include sillimanite, 
quartz, enstatite, and sapphirine. 

Polymorphism. Cordierites prepared hy- 
drothermally show different refractive in- 
dices from cordierites made under atmos- 
pheric pressure (Yoder, 1952). Whereas 
these dry cordierites have a mean refractive 
index (a + y)/2 of 1.522 + 0.003, all hydro- 
thermal cordierites made have a value of 
about 1.540 ±0.003. An index of refraction 
equal to that of the dry form was not ob- 
served in products prepared under water 
pressure. By heating hydrothermal cordier- 
ites under dry conditions for short periods 
of time only, the mean refractive index can 
be changed into the one of the dry form. 
The differences in mean refractive index of 
the cordierites are presumed to be due to 
variable amounts or lack of water. 

The structural change of orthorhombic 
cordierite to hexagonal indialite of the 
same composition (Miyashiro, 1957) is in- 
dependent of the above-mentioned change 
in mean refractive index. Both indialites 
and cordierites have been synthesized with 
high and low refractive indices. The crys- 
tals made at the lowest possible tempera- 
tures at 2000 bars and 5000 bars (550° to 
600° C) are high-index indialites or inter- 
mediate structural stages toward ortho- 
rhombic cordierites. It is believed that 
these indialites are metastable in this P-T 
range. In long runs at higher temperatures, 
and at all water pressures investigated, or- 
thorhombic cordierites were obtained up 
to the beginning of melting. 

Melting curve. The beginning of melt- 
ing of cordierite has been determined up 
to 10,000 bars water pressure. The P-T 
curve has the characteristic shape of other 
silicate melting curves, as shown by Yoder 
in last year's report. The slope of the curve 
is similar to that for plagioclase. At atmos- 
pheric pressure Mg cordierite melts incon- 
gruently to mullite + liquid at 1465° C. 
Under water pressure the melting remains 
incongruent but the products change. At 
2000 bars and about 1225° C cordierite be- 
gins to melt to mullite + spinel + liquid. At 
5000 bars and 1090° C it melts to spinel + 
sapphirine + liquid, and at 10,000 bars and 
960° C to sapphirine + liquid. Besides the 


incongruent melting products and a silica- 
rich vapor phase a variable amount of co- 
rundum is always present, apparently a 
leach product. These data indicate that for 
cordierite bulk composition the field of 
mullite + liquid is reduced at fairly low 
water pressures in the same way as shown 
for the muscovite composition (Yoder and 
Eugster, 1955) . They also indicate that the 
melting relations in the system MgO- 
AlaOa-SiC^-H-O deviate appreciably with 
increasing pressure from those in the dry 
system MgO-AUOa-SiCK. The mullite field 
is apparently reduced gradually by the ex- 
panding spinel and sapphirine fields. Runs 
on the incongruent melting of cordierite at 
various water pressures up to the liquid + 
vapor region are under way. Under dry 
conditions cordierite composition is all 
liquid above 1523° C. 

Solubility. The solubility of cordierite 
in water vapor was found to be incongru- 
ent at all water pressures and at tempera- 
tures above roughly 700° C. The solids 
dissolved by the vapor phase are therefore 
not of cordierite composition, but are usu- 
ally much richer in silica, or consist of 
silica alone. Some of the vapor phase was 
quenched as a siliceous glass either coating 
the crystalline charge or forming aggre- 
gates of clear balls. In addition, aluminous 
leach products like spinel, sapphirine, and 
corundum were formed in variable 
amounts depending on temperature, water 
pressure, and water content of the sealed 
platinum tube. The coexistence of cordier- 
ite and corundum does not necessarily im- 
ply that they form a stable assemblage. It 
may be recalled that corundum is found in 
metastable coexistence with quartz in some 

Synthesis of Fe cordierite. The hydro- 
thermal synthesis of Fe cordierite was 
achieved using a powder mix made of 
cristobalite, y-alumina, and Fe oxalate. It 
was not possible to prepare a glass of the 
requisite cordierite composition by the 
usual methods (nitrogen furnace, iron cru- 
cible), because its melting point is higher 
than that of metallic iron. In order to mini- 

mize oxidation the powder was sealed in 
gold tubes with excess water. No buffers 
were used. Under the hydrogen pressures 
generated by the reaction of water on the 
bomb wall, iron remained mainly in the 
ferrous state even in runs of several 
months' duration. The main difficulty of 
the synthesis is that FeO reacts rapidly 
with AI2O3 to form hercynite (possibly 
with a little magnetite in solid solution). 
The spinel in turn reacts only very slowly 
with Si0 2 to form Fe cordierite. This is 
especially true at elevated temperatures 
(650° C and higher at 2000 bars). At 2000 
bars, 500° C or lower, no Fe cordierite or 
hercynite was formed; the charge consisted 
of Fe chlorite, quartz, and some unreacted 
alumina. The best conditions for the 
growth of Fe cordierite were found to be 
at 2000 bars and 550° to 600° C. A 96-hour 
run at 600° C yielded about 75 per cent Fe 
cordierite in addition to quartz and hercy- 
nite. After 2 months about 90 per cent of 
the charge was Fe cordierite. At 5000 bars 
water pressure the formation of Fe cordier- 
ite is even slower, and at 10,000 bars Fe 
cordierite did not grow. A specimen of 
natural Fe-rich cordierite in intergrowth 
with quartz from Dosi, Japan (Shibata, 
1936), was made available for decomposi- 
tion experiments through the courtesy of 
Professor Shibata. The breakdown oc- 
curred mainly under the same P-T con- 
ditions as for the Mg cordierite and, in 
general, at a greater rate. It is concluded 
tentatively, therefore, that iron does not 
change the lower stability limits of cordier- 
ite appreciably. 

Geological application. The stability field 
of Mg cordierite as shown in figure 7 de- 
picts the maximum stability of the mineral 
under various temperatures and at water 
pressures equal to the total pressure. Intro- 
duction of iron into the structure does not 
seem to change the lower stability limits 
appreciably but does lower the melting 
curve. Other components occurring in 
natural rocks may influence the liquidus 
curve as well as the lower stability curve. 
In the presence of much K2O, for example, 



the mineral pyrophyllite would be replaced 
by muscovite so that the reaction to form 
cordierite would be: muscovite + chlorite + 
quartz—^ cordierite + phlogopite (biotite) . 
It is likely that this reaction would take 
place at higher temperatures than the chlo- 
rite + pyrophyllite reaction. 

Although the synthetic alkali-free system 
does not strictly apply to the much more 
complicated natural situation, a few impor- 
tant inferences concerning natural occur- 
rences of cordierites can be drawn. 

The lower stability curve shows that the 
formation of Mg cordierite in the presence 
of excess water vapor requires rather high 
temperatures (from 500° to 600° C). Cor- 
dierite, therefore, is to be expected as a 
common mineral in contact metamorphic 
aureoles. At water pressures less than the 
total pressure cordierite would be stable at 
lower temperatures. For this reason the 
width of the aureole would be determined 
by the rate of escape of water through frac- 
tures or adjoining porous strata. 

It has often been stated that cordierite 
does not occur in regional metamorphic 
rocks. Its occurrence in granulites (Lap- 
land; Eskola, 1952) or regional metamor- 
phic spotted slates (Banffshire, Scotland; 
Harker, 1939), and schists (Central Pyre- 
nees; Zwart, 1958), shows that this general- 
ization is not true. It is striking, however, 
that in most areas of progressive regional 
metamorphism cordierite has not been ob- 
served in any zones despite the presence of 
rocks of appropriate bulk composition. The 
experimental findings suggest that there 
are two main reasons for the absence of 
cordierite in these areas: the temperatures 
even in the highest zones of the area were 
below the lower stability limit of cordier- 
ite; and the pressures under which these 
rocks were formed, with or without water 
present, were higher than those at which 
cordierite breaks down into other hydrous 
or anhydrous phases. 

The mineral assemblages of the regional 
metamorphic rocks without cordierite give 
some clues about conditions under which 
they were formed. Aluminosilicates are of 
special importance: noncordierite-bearing 

progressive regional metamorphic rocks 
like those of Dalradian succession of the 
Scottish Highlands north of the Highland 
Boundary Fault do not contain andalusite, 
but only kyanite or sillimanite in progres- 
sively higher zones. Relating this observa- 
tion to the kyanite-sillimanite inversion 
curve as given by Clark, Robertson, and 
Birch (1957), one reaches the conclusion 
that these rocks must have formed under 
high confining pressures. Cordierite-kya- 
nite assemblages are extremely rare in na- 
ture, and most can be demonstrated to 
represent disequilibria (Suess, 1900; Scheu- 
mann, 1941, and personal communication, 
1959). This again sets a limit to the sta- 
bility field of cordierite. The absence of 
cordierite from most progressive regional 
metamorphic rocks, therefore, seems to be 
due to the prevalence of high pressures 
during their formation. 

Cordierite-bearing regional metamorphic 
rocks, on the other hand, very frequently 
contain andalusite. Cordierite-andalusite 
is, for instance, the typical assemblage of 
the spotted slates in Banffshire. In accord- 
ance with the commonly accepted low- 
pressure, low-temperature stability field of 
andalusite, these rocks must have formed 
under relatively low confining pressures 
and at temperatures just high enough to 
form cordierite. The assemblage cordier- 
ite-sillimanite is encountered in granulites 
and in aluminous gneisses, as for instance 
the ones from the Bavarian Forest (Fischer, 
1938) or Ellon, Aberdeenshire (Harker, 
1939). These rocks were probably formed 
at higher temperatures and possibly higher 
pressures than the andalusite-bearing rocks. 
Thus cordierite is in all probability a low- 
or medium-pressure mineral. The influence 
of shearing stress on the stability of cordier- 
ite, emphasized by Harker (1939) and Es- 
kola (1950), who describe it as an "anti- 
stress-mineral," has not been investigated. 


G. A. Chinner 

Studies of iron-bearing silicate minerals 
now being conducted in this laboratory are 


of significance in the interpretation of a 
sequence of regionally metamorphosed 
pelitic schists containing widely varying 
ferric/ferrous ratios which occur in Glen 
Clova, Angus, Scotland. 

Analytical work on these rocks and their 
enclosed minerals was commenced in the 
Department of Mineralogy and Petrology, 
University of Cambridge, and has been 
continued during the past year with the 
cooperation of Dr. E. G. Zies. Although 

tary origin, a conclusion supported by 
geochemical evidence to be published later. 
From a compositional viewpoint, however, 
these rocks may be treated as an essentially 
isofacial series of variable oxygen content. 
To express this variation in oxidation state 
the molecular percentage (2Fe 2 3 X 100)/ 
(2Fe 2 3 + FeO), termed the "oxidation 
ratio," has been used; rocks found in Glen 
Clova have oxidation ratios varying be- 
tween 6 and 75. 


1 o 

1 1 









■ — ^J{Je 





o ~^— . 















1 1 1 





.— —° 

o^— ■ ' 

— ~o 













1 lrono^f!___— 

-i — • i i 





20 30 40 50 

Oxidation ratio of rock 



Fig. 8. Modes of Glen Clova gneisses recalculated so that biotite + garnet + muscovite + opaque 
minerals = 100, plotted against the oxidation ratios of the respective rocks. 

the data obtained give information on sev- 
eral aspects of the metamorphic process, it 
is proposed here only to summarize the 
conclusions that are of interest with regard 
to present discussions on "closed" and 
"open" systems during metamorphism. 

The rocks under consideration are essen- 
tially quartz-oligoclase-muscovite-biotite- 
garnet-kyanite gneisses with opaque 
phases whose nature depends on the fer- 
rous/ferric ratio. The confinement of rocks 
of varying ferrous/ferric ratio to well de- 
fined sedimentary bands suggests these dif- 
ferences to be of premetamorphic, sedimen- 

In general, increasing oxidation ratios in 
these rocks are accompanied by increasing 
amounts of muscovite and iron oxides, and 
decreasing amounts of biotite and garnet 
(fig. 8). This amounts to a withdrawal of 
iron from the silicate minerals, probably 
due to the inability (in this environment) 
of biotite and garnet to accommodate sub- 
stantial amounts of ferric iron in their 
structures. The relationship between inter- 
bedded layers of varying oxygen content 
may thus be represented by the equation 
Fe ++ eastonite (biotite) +almandine + 0= 
muscovite + iron oxides + quartz. 


Since, however, in these rocks the mag- 
nesium content is contained virtually ex- 
clusively in biotite, and the manganese con- 
tent in garnet, diminishing amounts of 
these minerals with increasing ferric/fer- 
rous ratios in the rocks are accompanied by 
increasing MgO/FeO ratios in the biotites 
and increasing MnO/FeO ratios in the 
garnets (fig. 9). 

These rocks may be classified into three 
categories on the basis of the opaque min- 
erals present: ilmenite-magnetite-graphite- 

Oxidalion ratio of rock 

Fig. 9. MgO/(MgO + FeO) ratios of biotites 
and MnO/(MnO + FeO) ratios of garnets 
plotted against the oxidation ratios of their host 
rocks. Solid circles, minerals from ilmenite- 
magnetite-graphite assemblages; divided circles, 
minerals from ilmenite-magnetite-ilmenohema- 
tite assemblages; open circles, minerals from 
magnetite-ilmenohematite assemblages. 

bearing assemblages; ilmenite-magnetite-il- 
menohematite-bearing assemblages; and 
magnetite-ilmenohematite-bearing assem- 
blages. The ilmenite-magnetite-graphite- 
bearing rocks form a group of oxidation 
ratio 6 to 14, in which the carbon content 
was high enough to keep the ferric/ferrous 
ratio very low, and the partial pressures of 
oxygen constant. The compositions of bio- 
tite (45 per cent eastonite) and of garnets 
(2 per cent spessartite) in this group are 

constant to within the error of sampling 
and analysis. The second group, the ilmen- 
ite-magnetite-ilmenohematite-bearing as- 
semblage, is found in rocks of oxidation 
ratio of approximately 40. This was an in- 
variant assemblage in which, over a limited 
range of rock oxygen contents, the compo- 
sitions of the minerals, and hence the par- 
tial pressures of oxygen, were constant, 
variations in the total oxygen content being 
accommodated by varying proportions of 
magnetite, hematite, and ilmenite. 

The third group, the magnetite-ilmeno- 
hematite-bearing assemblages, compose 
rocks of oxidation ratio in excess of 40; in 
these the correlation of biotite MgO/FeO 
ratio and garnet MnO/FeO ratio with in- 
creasing rock oxidation ratio is found. For 
each rock oxidation ratio there is a spe- 
cific MgO/FeO ratio of biotite, a specific 
MnO/FeO ratio of garnet, and a specific 
titanium content of ilmenohematite. These 
are interpreted as having been univariant 
assemblages in which the oxygen content 
of the rock fixed the composition of the 
biotites and garnets, as well as the titanium 
content of the ilmenohematite, the mineral 
assemblage in its turn specifying the oxy- 
gen partial pressure of its immediate en- 
vironment. These observations thus sup- 
port the recent suggestions of Yoder (Year 
Book 56, p. 233) and Thompson (1957, 
p. 855) that during regional metamor- 
phism rocks in general behave as narrowly 
defined units closed to oxygen: the partial 
pressure of oxygen in each unit is deter- 
mined by the mineral assemblage, and 
hence the original oxygen content, rather 
than being imposed from outside through 
the agency of a vapor phase. 


The garnet group has attracted consid- 
erable attention since the recent discovery 
that some rare-earth synthetic members are 
ferrimagnetic. Knowledge gained from the 
synthesis of these new compounds has con- 
tributed greatly to an understanding of the 
role of minor elements in natural garnets. 
The major end members of the naturally 


occurring garnets have also been synthe- 
sized, but their interrelations have not been 
investigated. Because common garnet con- 
stitutes one of the most important minerals 
in metamorphic rocks, knowledge of their 
stability relations is required for interpret- 
ing the conditions of formation of these 
rocks. As in the investigation of any com- 
plex mineral group, it is necessary to deal 
first with the end members, then with their 
respective binary systems, ternary, and 
more involved joins. The common garnets 

Al 2 3 -Si0 2 . Since neither grossularite nor 
pyrope is stable at atmospheric pressure 
above approximately 850° C (the lowest 
temperature at which reaction rates are 
sufficiently rapid to allow crystallization 
from a glass), the join grossularite-pyrope 
is always quaternary. Because of the peno- 
logical significance of compositions along 
this join we have conducted quenching 
runs on glasses prepared at 10 weight per 
cent intervals to determine the subsolidus 
assemblages and temperatures of crystalli- 



SP + FO + CORD + AN 

GROSSULARITE 10 20 30 40 50 60 70 80 90 PYROPE 

3Ca0.AI 2 3 .3Si0 2 3MgO AI 2 3 .3Si0 2 

Weight per cent 

Fig. 10. Graphical presentation of data obtained on the join grossularite-pyrope. Abbreviations 
for solid phases as in figure 11; /, liquid. The inversion wollastonite-pseudowollastonite has been 

lie mainly in the almandite-pyrope-grossu- 
larite system of five components, portions 
of which have been under investigation. 
The results on the pyrope-grossularite join 
as well as some notes on the almandite- 
grossularite join under both anhydrous and 
hydrous conditions are presented below. 

The Join Grossularite-Pyrope at Atmos- 
pheric Pressure 
G. A. Chinner and J. F. Schairer 
Members of the grossularite (3CaO* 
Al 2 CV3Si0 2 ) and pyrope (3MgO • A1 2 3 ■ 
3Si0 2 ) series lie in the system CaO-MgO- 

zation of the constituent minerals at 1 atm 
pressure. The results, which serve as a basis 
for the more complex hydrothermal work 
on the same compositions, are presented in 
the projection on figure 10. Moreover, by 
combining our data with those of previous 
workers in the quaternary system CaO- 
MgO-Al 2 3 -Si0 2 , we have been able to 
deduce the relationship of the quaternary 
invariant points and univariant lines in the 
important silica-rich portion of this sys- 
tem: this is shown diagrammatically in fig- 
ure 11. 
Our data give information on the tern- 



peratures of four of these invariant points — 
H, I, J, and K. Compositions lying between 
grossularite and approximately Gro7sPy22 
are represented in the subsolidus region by 
the assemblage pseudowollastonite-pyrox- 
ene-melilite-anorthite, which melts at 
1235° ± 5° C, the temperature of the qua- 
ternary eutectic K. Between approximately 
Gro7sPy22 and Gro4sPy55 the subsolidus 

spinel melts at the temperature of this 
reaction point. Between Gr03ePye4 and py- 
rope the assemblage anorthite-forsterite- 
spinel-cordierite melts at 1280° ± 5° C, the 
temperature of the quaternary eutectic H. 
Along the univariant lines between the 
three quaternary eutectics H, J, and K 
lie two important temperature maxima. 
Thus the univariant anorthite + forsterite + 

(CaO-MqO-Si0 2 ) tr 

(MgO-AI 2 3 -Si0 2 XCaO-MgO-Si0 2 ) ,r (Co0-AI 2 O 3 -SiO 2 ) 

1470 ±10° py,ox 1470 1 10° 1470 i 10° 


(CoO-MgO-Si0 2 ) 

(Ca0-Al,0,-Si0,) P wo1 

1165 + 5° ,r 


(Co0-Al 2 3 -Si0 2 ) 
1265 + 5° 


on > J f 
pwol ? 



(MgO-AI 2 OySi0 2 )_ c S P rd 

(Mg0-Al 2 3 -Si0 2 ) f (CaO-AI 2 3 -Si02) 

1355 + 5° 



(Mg0-AI 2 3 -Sj0_ 2 , cor 
1360 + 5° 










CORD t '° 


pyo^y j 

MEL X A ■ 





TR k 

(Mg0-Al 2 3 -Si0 2 ) 
1460 + 5° 

" S °P spin (Mg0-Al 2 3 -Si0 2 ) 
> m < mu — »C 
sop / A sap |4 8 2 + 5° 

h (CaO-AI 2 3 -Si0 2 ) 





I440±5° AN 

(MgO-AI 2 3 -Si0 2 ) an 
^ fo 





1357 + 5° 





1512 + 5° 

co (Mg0-Al 2 3 -Si0 2 ) 

j t < mu »i 

y B sp 1575+5° 


T fo 




(CoO-MgO-Si0 2 ) 

Fig. 11. The relation of ternary invariant points (small solid circles, lettered a to v) in the limit- 
ing systems to univariant lines and quaternary invariant points (large solid circles, lettered A to K) 
within the silica-rich portion of the system CaO-MgO-Al 2 3 -Si0 2 . This diagram merely depicts 
phase relationships. The lines and points do not lie on a plane; their spatial angular relationships 
as shown, the length of lines, and the position of temperature maxima on these lines are arbitrary 
and without significance. Arrows indicate directions of falling temperature. Abbreviations for solid 
phases: tr, tridymite; cr, cristobalite; mu, mullite; co, corundum; pyrox, pyroxene; sp, spinel; cord, 
cordierite; sap, sapphirine; fo, forsterite; an, anorthite; mel, melilite; pwol, pseudowollastonite. 

assemblage pyroxene-melilite-anorthite- 
forsterite melts at 1225° ± 5° C, the tem- 
perature of the quaternary eutectic /; the 
proportion of pyroxene in this is very small 
and cannot always be identified either by 
optical or by X-ray means. At 1238° 
±5° C these compositions pass through 
the quaternary reaction point /, melilite- 
anorthite-forsterite-spinel-liquid. In a very 
narrow range of compositions between 
Gro45Py55 and Gro 3 cPy64 the subsolidus 
assemblage melilite-anorthite-forsterite- 

spinel + liquid will crystallize at the qua- 
ternary H (anorthite-forsterite-spinel-cor- 
dierite-liquid) if its composition lies to 
the pyrope side of the anorthite-forster- 
ite-spinel plane, but will crystallize at 
either / (melilite-anorthite-forsterite-spi- 
nel-liquid) or / (melilite-anorthite-forster- 
ite-pyroxene-liquid) if its composition lies 
to the grossularite side of that plane. Simi- 
larly the univariant pyroxene + anorthite + 
melilite + liquid will crystallize at / if its 
composition lies to the pyrope side of the 



anorthite-melilite-pyroxene plane, and to 
K (pseudowollastonite-melilite-pyroxene- 
anorthite-liquid) if its composition lies to 
the grossularite side of that plane. These 
two planes are therefore ternary systems 
and separate significant volumes in this 
part of the system CaO-MgO-Al 2 3 - 
SiO a . It is also important to note that, since 
five phases are represented at each of the 
quaternary points found for compositions 

metasediments, their first appearance be- 
ing marked by the garnet isograd. Succes- 
sive reactions involve garnet in the produc- 
tion of other critical phases. Petrographic 
affinities of garnets of varying compositions 
are indicated in figure 12. In general, there 
is complete solid solution between alman- 
dite and pyrope and limited solubility be- 
tween grossularite and both almandite 
and pyrope. The area in which no nat- 

Androdile * Spessortile 

Calcareous Rocks 


Fig. 12. Plot of natural garnets low in andradite and spessartite from various rock types (mainly 
from Wright, 1938). 

along this join, the melilites involved must 
be of a definite and fixed composition. 

Grossularite- Aim an dite-Pyrope- Water 


H. S. Yoder, Jr., and G. A. Chinner 

Almandite (Fe 3 Al 2 (Si04)3), grossular- 
ite (Ca 3 Al 2 (Si04) 3 ), and pyrope (Mg 3 Al 2 - 
(Si0 4 ) 3 ) are the predominant end mem- 
bers of the common garnets found in the 
metamorphic rocks and in some igneous 
rocks. Garnets are important products in 
the progressive metamorphism of pelitic 

ural garnets are known is a region of 
composition in which two garnets of maxi- 
mum solid solution would be expected to 
exist. No rocks are known which contain 
two common garnets of the anticipated 
compositions. 2 As may be seen in figure 6, 
the area between the limits of the two 
garnet solid solutions is occupied in one 
case by either the assemblage Ca garnet + 

2 The two-garnet rock described by Winchell 
(1947) was later shown to consist of garnet + 
spinel (personal communication from H. Win- 
chell, 1958). 



plagioclase + clinopyroxene, plagioclase + 
kyanite + clinopyroxene, or Fe-Mg garnet + 
kyanite + clinopyroxene. It was the intent 
of the present study to outline the exact 
limits of solubility of these garnets, deter- 
mine the extent of their stability at various 
temperatures and pressures, and prepare 
determinative curves based on the proper- 
ties of the synthetic products. This am- 

nets along the joins grossularite-pyrope 
and grossularite-almandite. Since both 
grossularite and almandite grow within 
minutes at 10 kilobars water pressure, the 
section at this pressure was chosen for pre- 
liminary study. 

Grossularite-pyrope-water. Some eighty 
runs on the join grossularite-pyrope-water, 
a plane in a five-component system, reveal 




Mol per cent 

Fig. 13. Relationship of the anorthite (An) + diopside (Di) + spinel (Sp) plane to the grossu- 
larite (Gr)-pyrope (Py) join in a projection of the CaO-MgO-Al 2 3 -Si0 2 -H 2 system. End- 
member compositions of other phases encountered in the hydrous as well as the anhydrous study are 
indicated as follows: Geh, gehlenite; Ak, akermanite; Wo, wollastonite; Tr, tremolite; Ts, tscher- 
makite; Co, cordierite; Ged, gedrite; Ams, amesite; CI, clinochlore; Anth, anthophyllite; En, ensta- 
tite; Serp, serpentine; Fo, forsterite. 

bitious plan was immediately beset with 
great complexity and difficulties. 

The upper stability curves for almandite 
(Yoder, 1955) and grossularite (Yoder, 
1950£ and unpublished data) had been de- 
termined up to 10 kilobars; pyrope, how- 
ever, had been produced by Coes and 
others only at high temperatures and pres- 
sures between 20 and 30 kilobars. The first 
objective, therefore, was to synthesize gar- 

a very complex set of assemblages. The 
join anorthite + diopside + spinel + water 
uniquely divides the plane of interest into 
two portions midway (see fig. 13); the 
anhydrous composition at the intersection 
may be written alternatively as 

Ca 3 Al2Si 3 Oi2 + Mg3Al 2 Si 3 Oi2= 

grossularite pyrope 

CaAl 2 Si 2 Os + MgAl 2 4 + 2CaMgSi 2 O c 

anorthite spinel diopside 


Below the solidus, the calcium-rich part 
of the plane consists of a small field of 
grossularite solid solution + vapor, the gar- 
net containing less than 10 per cent by 
weight of pyrope, and a large region of 
grossularite solid solution + diopside + an- 
orthite + spinel + vapor. The liquidus has 
a small field of grossularite, which melts 
congruently at this pressure. Spinel is the 
first phase to appear on the liquidus over 
the remainder of the plane. 

Compositions in the magnesium-rich 
part of the plane crystallize with a variety 
of phases. Immediately below the liquidus 
such minerals as spinel, forsterite, diopside, 
anorthite, and an aluminous enstatite ap- 
pear. At lower temperatures an amphibole 
appears in the liquid at the expense of 
some of the above-named phases. At still 
lower temperatures (~825° C) chlorite is 
observed in the most magnesium-rich com- 
positions. The relations are greatly ob- 
scured by quench products, which include 
diopside, aluminous enstatite, montmoril- 
lonite, an amphibole, and chlorite. The 
present data suggest that the magnesium- 
rich part of the system is probably subdi- 
vided on complete crystallization into two 
volumes, one comprised of diopside + an- 
orthite + amphibole + spinel + vapor, the 
other of amphibole + chlorite + anorthite + 
spinel + vapor. Runs on synthetic pyrope 
prepared by L. Coes and F. R. Boyd indi- 
cated that the magnesium garnet is not 
stable at 10 kilobars water pressure. At the 
same temperatures, pressure, and duration, 
experiments on glass of pyrope composi- 
tion yielded inconsistent results. With 
changes in temperature and water pressure 
the appearance of aluminous enstatite, 
gedrite, and chlorite indicates important 
changes in mineralogy. The assemblages 
in the grossularite-pyrope system stable in 
the absence of water at atmospheric pres- 
sure have been presented in another section 
(fig. 10) . The phases wollastonite, melilite, 
and cordierite, observed in the anhydrous 
experiments, are absent from the assem- 
blages formed in the hydrous experiments. 
Their absence indicates again the signifi- 

cant changes in assemblages induced by 
water pressure and temperature. 

Grossularite-almandite-water. Investiga- 
tion of the grossularite-almandite-water 
plane is dependent primarily on maintain- 
ing the bulk composition of the system. 
That is to say, free oxygen must be ex- 
cluded to prevent the formation of solid 
solutions involving predominantly andra- 
dite, Ca 3 Fe2 +3 Si 3 Oi2, and possibly Fe3 +2 - 
Fe 2 +3 Si30i2 in addition to magnetite. Al- 
mandite had previously been synthesized 
using glass or an oxalate-bearing mixture 
in short runs at high pressures. The crys- 
tals, exceptionally well formed and green, 
were not analyzed for the state of oxidation 
of the iron. Runs on a specific bulk com- 
position in the system using this technique 
as well as the wiistite-magnetite buffer 
(Eugster, 1957) have now been chemically 
analyzed and compared. In half-hour runs 
at 950° C and 10 kilobars using gold tubes 
the products are less oxidized than those in 
buffered platinum tubes. In 1- and 2-hour 
runs using gold tubes oxidation slowly in- 
creases, whereas in buffered platinum tubes 
the oxidation state remains relatively con- 
stant. In these runs, consequently, the oxi- 
dation state of the products is somewhat 
higher in the gold tube than in the buf- 
fered platinum tube. These preliminary re- 
sults indicate that it will be possible to 
outline the limits of solid solution in the 
grossularite-almandite-water system in 
short runs using gold tubes. The curves 
derived will be tested with long buffered 

Geological application. The initial suc- 
cess in the elucidation of the complex gros- 
sularite-almandite-pyrope system indicates 
relationships of greater significance than 
had previously been anticipated. The sys- 
tem, although in itself deficient in silica, is 
represented in most of the volumes of the 
system CaO-Al 2 3 -MgO-Si0 2 which are 
important for rocks low in K 2 0. It indi- 
cates the significance of the anorthite-di- 
opside-spinel plane in separating the essen- 
tially anhydrous calcium-rich assemblages 
from the hydrous magnesium-rich assem- 


blages. The present results again emphasize 
the need for defining the critical fades in 
terms of coexisting assemblages rather than 
on the basis of a single assemblage. For 
example, the coexistence of amphibole and 
chlorite is usually interpreted as evidence 
of low-grade regional metamorphism, yet 
these minerals may coexist over a wide 
range of experimental conditions. Amphi- 
bole + chlorite is indeed a low-grade assem- 
blage if contiguous rocks contain assem- 
blages such as biotite + chlorite + muscovite, 
epidote + chlorite, or talc + amphibole. On 
the other hand, amphibole + chlorite may 
be a high-grade assemblage if the contigu- 
ous rocks of different bulk composition 
contain pyroxene + plagioclase or garnet + 
plagioclase + amphibole. Since every bulk 
composition may be represented at every 
pressure and temperature, and not all as- 
semblages are equally sensitive to changes 
of pressure and temperature, the aggregate 
of assemblages must be considered before 
the facies is uniquely defined. 

Garnet-Cordierite Parageneses 
G. A. Chinner 

Although commonly developed in the 
more aluminous schists at the higher 
grades of regional metamorphism, alman- 
dine garnet with low spessartite and gros- 
sularite contents is generally unstable 
under the lower pressures and higher tem- 
peratures of thermal metamorphism. Many 
instances have been described, of alman- 
dines formed during a period of regional 
metamorphism, which have completely or 
partly decomposed when subsequently 
remetamorphosed within the aureole of an 
igneous intrusion. More rarely, however, 
almandine garnets are seen to have existed 
quite stably in such environments. 

Where garnetiferous beds of the region- 
ally metamorphosed Scottish Dalradian se- 
quence have been remetamorphosed within 
the aureole of the Glen Doll intrusive com- 
plex, two modes of garnet behavior may be 
discerned : in certain bands garnet of mo- 
lecular composition AlmsiPynSp4Gro3- 

Andi is found to have coexisted stably with 
cordierite and spinel, whereas in neighbor- 
ing bands regional garnet of similar com- 
position (Alm8oPyiiSp 5 Gro2And 2 ) is seen 
to have been decomposing to cordierite- 
bearing aggregates. Such altered garnet, 
however, appears to have been involved 
with surrounding biotite by a reaction of 
the type Mg biotite + garnet "^Fe biotite -f 
cordierite; garnets armored from reaction 
by immersion in an inert mineral such as 
quartz or feldspar show no evidence of 

These observations suggest that the al- 
mandine was, in fact, not breaking down 
because its physical stability field, as de- 
fined by Yoder (1955), had been exceeded, 
but because it could no longer exist in its 
chemical environment. This chemical en- 
vironment was, however, well suited to its 
development under the P-T conditions of 
regional metamorphism; the suppression 
of garnet within the aureole was due to a 
change in the compositional field in which 
it could occur, consequent upon the forma- 
tion during thermal metamorphism of 
the newly stable phase, cordierite. To in- 
terpret the behavior of garnet within the 
thermal aureole, therefore, it is necessary to 
determine how cordierite modifies or re^ 
duces the compositional field within which 
almandine may form a stable phase. Eskola 
had suggested in 1915 that a limit exists in 
the amount of iron that can be accommo- 
dated by the magnesian molecule of cordi- 
erite, and that almandine appears in cordi- 
erite-bearing assemblages when, owing to 
a high FeO/MgO ratio in the rock, this 
limit is exceeded; data from the Glen Doll 
complex give support for such a hypothesis. 
Bands containing stable garnet (with cor- 
dierite) are richer in iron than associated 
bands in which regional garnet has partly 
disintegrated, and while a wide composi- 
tional range of cordierites occurs in differ- 
ent hornfelses, those cordierites occurring 
with almandine have the highest FeO/ 
MgO ratio. Similarly a cordierite separated 
from a disintegrating regional almandine 


is not, as might be expected, almost pure 
iron cordierite, but has a composition very 
similar to that occurring in hornfelses bear- 
ing stable almandine. It would thus appear 
that this composition (Mg46Fe S 4) repre- 
sents the maximum content of iron that (in 
the Glen Doll aureole) cordierite could 
accommodate. In terms of a triangular dia- 
gram with AI2O3, FeO, and MgO at the 
apices (fig. 14), rock compositions of in- 
creasing FeO/MgO ratio can be repre- 
sented by tie lines between biotite and cor- 
dierite of increasing FeO/MgO ratio, until 
at the limiting iron content of the cordierite 

A=AI 2 3 -(K z + Na 2 0t2Ca0) 


and it is thus clear that, if the iron content 
of cordierites is indeed limited, the limit 
itself must vary with the physical condi- 
tions of the environment. This raises the 
possibility of using the Fe/Mg ratio of 
cordierite in equilibrium with garnet as an 
indicator of the physical conditions of the 
environment in which it crystallized. 

In general, the FeO/MgO ratios of nat- 
ural cordierites show a relation to the in- 
ferred pressures of the environments in 
which they are found : cordierites occurring 
with garnets in regionally metamorphosed 
rocks of the granulite facies are of lower 


Fig. 14. 

(a) M=MgO F (b) M F (c) M 

AFM diagrams (molecular per cent) to show the relationship between (a) regional, 
cordierite-free assemblages and (b) thermal, cordierite-bearing assemblages in Glen Clova. Open 
circles represent analyzed phases, (c) demonstrates the postulated direction of change of the garnet- 
biotite-cordierite triangle, and the consequent restriction of the field of rock compositions in which 
garnet can occur, as the maximum iron content of cordierite is increased by either decreasing pres- 
sure or a combination of decreasing pressure and increasing temperature. 

the composition lies within the three-phase 
triangle cordierite-biotite-garnet. 

The paucity of almandine garnets in this 
and in other thermal aureoles may thus be 
due to the rarity of pelitic rocks with effec- 
tive FeO/MgO ratios high enough to lie 
within this three-phase triangle. 

Furthermore, unless the range of compo- 
sitions in the garnet-cordierite two-phase 
field is excessively wide, the Eskola concept 
of a fixed limiting iron content of cordier- 
ite implies that cordierites in equilibrium 
with garnet should always be of this 

Analyzed cordierites from garnetiferous 
rocks in other localities, however, have 
widely differing compositions (fig. 15), 

FeO/MgO ratio than cordierites from gar- 
netiferous xenoliths in igneous intrusions, 
and the most iron rich cordierites recorded 
from nature (Shibata, 1936) occur in a peg- 
matite. This relationship would suggest 
that increase in pressure reduces the 
amount of iron which the cordierite struc- 
ture can accommodate, but there would 
presumably also be a temperature effect of 
greater or lesser magnitude acting in oppo- 
sition to pressure. 


/. F. Schairer and N. Morimoto 

During the past year substantial progress 
has been made on this system, which de- 


Cordiente d- 


Fe - Cordierite 



3AI 2 3 7Si0 2 


- + —A AT 


>Fe Cordierite 

\ \ 

\ \ 

(i \ \ 

_ onol" ^ "Thermal .Pegmatitic 

X x assemblages \assemblage^ gssemb | age \ 







Fig. 15. The cordierite-Fe cordicrite-almandite-pyrope plane showing (a) tie lines between ana- 
lyzed cordierite-garnet pairs and (b) the compositions of analyzed cordierites and garnets from 
garnet cordierite-biotite-feldspar (± spinel, pyroxene, quartz) rocks of regional (open circles), 
thermal (solid circles), and pegmatitic (cross) origin. It is probable that the minimum pressures 
at which pyrope-rich garnets can exist are greater than the maximum pressures of cordierite sta- 
bility. Compositions in the magnesium-rich portion of this diagram will thus be represented not 
by garnet-cordierite pairs but by combinations either of cordierite and other phases (e.g., pyroxene, 
sapphirine, spinel) or of garnet and other phases. 


picts the mutual melting relations between 
certain olivines, pyroxenes, and feldspars. 
Last year (Year Book 57, pp. 212-213) we 
discussed the importance of this system in 
experimental petrology and gave a pre- 
liminary phase-equilibrium diagram for 
one of the limiting systems (forsterite-al- 
bite-diopside). Now we have nearly com- 
pleted the join protoenstatite-diopside- 
albite, which roughly bisects the tetrahe- 

der pattern methods. The phase-equilib- 
rium diagram for this join is given here as 
figure 16. An examination of it shows that 
the join cuts the primary phase volumes of 
forsterite, diopsidic pyroxenes, and soda- 
rich plagioclases at liquidus temperatures. 
The interesting relations at temperatures 
below liquidus temperatures can only be 
shown with the aid of an extensive series 
of isothermal planes. We have selected the 

• 1148° 

-1130 ±2" 

MgO.Si0 2 


Fig. 16. Equilibrium diagram of the join protoenstatite-albite-diopside showing compositions 
studied, primary phase volumes cut by this join, the piercing point of a quaternary univariant line, 
and isotherms. 

dron representing the system forsterite-di- 
opside-silica-albite. This is a most interest- 
ing join because it portrays the relations be- 
tween the olivine forsterite, a pyroxene or 
two conjugate pyroxenes, soda-rich plagio- 
clase feldspars, and liquid. 

Seventy-nine separate compositions in 
the join protoenstatite-diopside-albite were 
prepared and studied by the method of 
quenching. The solid phases present at 
various temperatures between liquidus 
temperatures and 1020° C were deter- 
mined by either microscopic or X-ray pow- 

isotherms for 1350°, 1300°, 1250°, 1200°, 
and 1150° C, which are presented here as 
figures 17 through 21, respectively. When- 
ever forsterite crystals are present, the com- 
position of the liquid phase cannot lie in 
the join. Similarly, if any of the pyroxene 
crystals contain AUO3, i.e., if they are not 
binary solid solutions of MgSiOs and 
CaMg(SiOs)2, the composition of the liq- 
uid phase cannot lie in the join. Compari- 
son of X-ray powder patterns of diop- 
sidic pyroxenes present in compositions in 
the join protoenstatite-diopside-albite with 




No 2 AI 2 3 .6Si0 2 


MgO Si0 2 



CaO.MgO 2S.0 2 

Fig. 17. Isotherm for 1350° C for the join protoenstatite-albite-diopside. The compositions of 
the liquid phase do not lie in this join if forsterite is one of the solid phases or if the pyroxene crys- 
tals contain A1 2 3 . In this figure and in the four subsequent figures the following abbreviations are 
used: Fo, forsterite; Proto, protoenstatite; Diop, diopsidic pyroxenes; Liq, liquid. 


Na 2 O.AI 2 3 .6Si0 2 


MgO.Si0 2 


Fig. 18. Isotherm for 1300° C. 



No,0 Al 2 3 6Si0 2 


MgO.Si0 2 

10 20 30 40 SO 60 70 


Fig. 19. Isotherm for 1250° C. 

CaO.MgO.2Si0 2 


No 2 AI 2 Oj 6Si0 2 


MgO.Si0 2 

10 20 30 40 SO 60 70 SO 90 


Fig. 20. Isotherm for 1200° C. 

CoO.MgO,2Si0 2 


those in the system protoenstatite-diopside 
(where no AI2O3 is present in the composi- 
tion) suggests that the AI2O3 content of 
the diopsidic pyroxenes present in mixtures 
in the join must be very low or that the 
presence of AI2O3 does not appreciably 
change the X-ray spacings. 

Boyd and Schairer (Year Book 56, pp. 
223-225) were able to determine the com- 
position of MgSi0 3 -rich protoenstatites in 

attempt to ascertain precise composition 
data on these solid solutions. 

P. M. Orville 

The subsolvus relations in the NaAl- 
SisOs-KAlSisOs feldspar system are of par- 
ticular interest to petrologists because most 
perthites and cryptoperthites which are un- 
mixed into Na- and K-rich feldspar phases 


NOjO.AI^Oj 6Si0 2 


MgO.Si0 2 


Ca0.Mg0.2SiO 2 

Fig. 21. Isotherm for 1150° C. 

the system MgSi0 3 -CaMg(Si0 3 )2 from 
the spacings of clinoenstatites formed from 
the protoenstatites on cooling. So far we 
have not been able to utilize this technique 
for protoenstatites in the join protoensta- 
tite-diopside-albite because the protoensta- 
tites on cooling always yield a mixture of 
protoenstatites, clinoenstatites, and ortho- 
enstatites in various proportions depending 
on the rate of cooling. Such a mixture has 
proved unsuitable for estimation of com- 
position from X-ray spacings. During the 
coming year we hope to make X-ray meas- 
urements at elevated temperatures in an 

contain only minor amounts of other feld- 
spar molecules and closely approach the 
synthetic alkali feldspar system in com- 
position. Tuttle and Bowen (1958) and 
Yoder, Stewart, and Smith (Year Book 56) 
have studied the subsolvus unmixing rela- 
tions in the synthetic alkali feldspar system 
in the presence of H2O vapor, but because 
of the sluggishness of reaction they were 
not able to demonstrate that the solvus 
obtained in this way represented a close 
approach to equilibrium. 

A new approach to feldspar reactions 
in the solvus region is suggested by experi- 



ments of Wyart and Sabatier (1956), who 
found that natural Na feldspar can be con- 
verted to K feldspar, and natural K feld- 
spar to Na feldspar, in the presence of 
alkali chloride water solutions over a con- 
siderable range of temperature and pres- 
sure. Experiments carried out during the 
past year demonstrate that reaction rates 
in the solvus region of the alkali feldspar 

700° C, 2000 bars 

NaAISi 3 8 

Fig. 22. Alkali feldspar — 2 molar alkali chlo- 
ride solution system at 700° C and 2000 bars 
total pressure. Experimentally determined tie 
lines for 2-day runs. Alkali feldspar composition 
determined by 201 X-ray method. Solution com- 
position determined by flame photometer. 

system are accelerated in the presence of 
alkali ions in solution. 

In the experiments so far completed the 
charge consists of an alkali chloride solu- 
tion (most commonly 2 molar) and a syn- 
thetic feldspar which has been crystallized 
hydrothermally from a glass. Weighed 
amounts of these materials are placed in a 
sealed platinum capsule. All runs have 
been made at 2000 bars ELO pressure and 
300° to 700° C, in periods of 1 to 10 days. 
Under these conditions the material within 
the tube consists of a water-rich salt solu- 

tion and one or two crystalline feldspar 
phases. Upon completion of a run the 
compositions of the feldspar phase or 
phases are determined by the 201 X-ray 
method (Orville, Year Book 57; Bowen 
and Tuttle, 1950). The composition of the 
solution is determined by the flame pho- 

KAISi 3 8 

0,2000 bars 

NaAISi 3 8 

Fig. 23. Alkali feldspar — 2 molar alkali 
chloride solution system at 600° C and 2000 bars 
total pressure. Experimentally determined tie 
lines for runs of 4 to 10 days. Alkali feldspar 
composition determined by 201 X-ray method. 
Solution composition determined by flame pho- 

Let us consider the experimental results 
obtained at 2000 bars pressure and two 
different temperatures, one above the crest 
of the alkali feldspar solvus and one below. 
Figure 22 shows the results at 700° C, the 
higher temperature. One feldspar plus 
alkali chloride solution is the only phase 
assemblage present for all bulk composi- 
tions, and a series of tie lines connects each 
feldspar composition with a solution com- 
position. Equilibrium is closely approached 
within 24 hours for any bulk composition. 



The same bulk composition gives identical 
tie lines regardless of the compositions of 
the feldspar and alkali chloride solution 
used as starting materials. Figure 23 shows 
results at 600° C, below the crest of the 
solvus. A three-phase invariant assemblage 
consisting of two feldspars plus alkali 
chloride solution is bounded by assem- 
blages of one feldspar plus alkali chlo- 
ride solution. Equilibrium is closely ap- 
proached within 4 days over the greater 
part of the one-feldspar regions at this 
temperature. Equilibrium is less quickly 
approached within the two-feldspar region, 
but it has been possible to determine the 
Or-rich side of the alkali feldspar solvus to 
within ±3 weight per cent down to 
400° C. 

It can be seen in figure 23 that the alkali 
chloride solution coexisting with two alkali 
feldspar phases at 600° C is poor in K and 
rich in Na. At lower temperatures the 
proportion of K falls still lower. At 600° C 
the mole per cent K/(Na + K) in the solu- 
tion is 0.236, at 500° C it is 0.198, and at 
400° C it is 0.163. 

Although figures 22 and 23 are based on 
experiments with solutions 2 molar with 
respect to alkali chloride, other concentra- 
tions have been used with very nearly the 
same results. With a total concentration 
of alkali chlorides of 0.2 molar, a K/(Na + 
K) ratio of 0.242 was obtained under the 
conditions described above in which a 2.0 
molar concentration yielded a ratio of 0.236. 

No crystalline phases other than Na-rich 
and K-rich feldspar have been observed in 
the final products of the runs. Since the 
total molar quantity of alkalies in the solu- 
tion is the same at the conclusion of a run 
as at the beginning, there can be no sig- 
nificant reaction between the feldspar and 
alkali chloride solution except ion ex- 
change of Na and K. The lower tempera- 
ture limit at which this reaction takes place 
at a significant rate has not been deter- 
mined, but it is less than 300° C. Mac- 
Kenzie (1957) found that glasses of albite 
composition crystallized as analcite at tem- 
peratures of 400° C or less at 2000 bars 

H 2 pressure. It is possible that analcite 
crystallized from albite glass under these 
conditions may form metastably, and the 
fact that the starting material in the present 
experiments is crystalline albite explains 
the absence of analcite. On the other hand, 
if analcite of hydrated albite composition 
is stable under these conditions its failure 
to appear may be due to the sluggishness 
of the reactions. 

It has been found that with certain modi- 
fications the methods used to study the 
alkali feldspar system can be applied to the 
ternary feldspar system. A major draw- 
back to ordinary hydrothermal research in 
the ternary feldspar system has been the 
impossibility of determining the ternary 
composition of a feldspar by an X-ray 
method. Through the use of alkali chlo- 
ride solutions the Na and K contents of a 
ternary feldspar can be varied while Ca 
remains constant, since Ca does not ex- 
change with alkalies in the solution under 
the conditions of the experiment. Within 
the one-feldspar region the ternary com- 
position of a feldspar can be determined on 
the basis of the known starting composi- 
tion and the Or content of the final product 
as determined by the 201 X-ray method. 
Precise location of the ternary solvus field 
boundary at a given temperature and de- 
termination of the alkali solutions coexist- 
ing with ternary feldspars on that solvus 
could determine the tie lines for coexisting 
ternary feldspars. 

This same general ion-exchange method 
may be applicable to study of other mineral 
systems which contain Na and K end 
members, for example, muscovite-parago- 
nite and nepheline-kaliophilite. It should 
be especially useful for determining the 
limits of solid solution when the amount 
of solid solution is very small. Since the 
solution end of the tie line in an isothermal 
section must swing from a pure Na or K 
salt solution for the pure end members to 
the equilibrium ratio for coexistence with 
two crystalline phases, the composition of 
the solution should be a very sensitive 
indicator of whether the limit of solid solu- 


tion has been exceeded or not. Small 
changes in the composition of a crystalline 
phase can be readily determined, owing to 
the precision with which the starting and 
final composition of the solution can be 

The moderate temperature and high rate 
at which ion-exchange reactions between 
an alkali salt solution and feldspars take 
place in the laboratory, together with the 
common occurrence of concentrated alkali 
salt solutions within fluid inclusions in 
natural minerals, suggest that such reac- 
tions may play an important role in nature. 

A number of investigators have shown 
that fluid inclusions from quartz and other 
minerals commonly contain 2 to 5 molar 
salt solutions. Some of the solutions were 
supersaturated, and cubic crystals, presum- 
ably of alkali chloride, were observed. All 
quantitative determinations of the non- 
volatile fraction of fluid inclusions have 
shown the presence of significant amounts 
of alkali salts, predominantly as chlorides, 
carbonates, and sulfates. The nearly ubiq- 
uitous occurrence of fluid inclusions in 
minerals of igneous and metamorphic 
rocks indicates that crystallization of the 
rocks has commonly taken place in the 
presence of a vapor phase containing alkali 
salts in solution. Investigation of the ion- 
exchange relations between alkali-bearing 
solutions and one or more feldspar phases 
may shed light on the transfer of alkalies 
during metamorphism and the late stages 
of magmatic crystallization. 

The alkali ratio of fluid inclusions in 
quartz should indicate whether the solu- 
tions were in equilibrium with the asso- 
ciated feldspar phases. It would be par- 
ticularly interesting to know this ratio for 
concentrically zoned pegmatites in which 
extreme segregation into monomineralic 
zones has commonly taken place. The 
problem is currently being investigated in 
a cooperative program with Dr. Edwin 
Roedder of the U. S. Geological Survey. 

Preliminary results on the ternary feld- 
spar compositions with the alkali chloride 
solution technique show that, the more 

calcic the plagioclase coexisting with an 
alkali chloride solution, the higher the 
proportion of K in the solution. Two 
ternary feldspar pairs which have different 
proportions of anorthite in the plagioclase 
feldspar and which are in close association 
with the same alkali-bearing solution can 
approach equilibrium with the solution 
and each other by exchange of K and Na 
ions. K ions from the calcic plagioclase- 
alkali feldspar pair go to the less calcic 
plagioclase-alkali feldspar pair, and Na 
ions from the less calcic plagioclase-alkali 
feldspar pair go to the calcic plagioclase- 
alkali feldspar pair. The process is very 
likely to take place under conditions of 
regional metamorphism in which rocks of 
heterogeneous composition are subjected 
to elevated temperature and pressure in the 
presence of a fluid phase. Operating under 
such conditions it provides a driving force 
for movement of K and Na. Its effect 
would be to deplete calcium-rich rocks and 
enrich calcium-poor rocks in K, with the 
reverse true for Na. 

W. G. Ernst 

Alkali amphiboles occur in a wide range 
of lithologic environments, from authi- 
genic and low-grade metamorphic to mag- 
matic. Riebeckite-arfvedsonite and glau- 
cophane are the most common members 
of this mineral group. The P-T stability 
ranges of magnesioriebeckite and glauco- 
phane were presented in Year Books 56 
and 57. Riebeckite-arfvedsonite, the char- 
acteristic amphibole of alkalic granites and 
syenites, has been synthesized, and pre- 
liminary results are now available. Investi- 
gation of glaucophane has been continued 
along three lines: a new petrologic treat- 
ment of glaucophane schists based on 
graphical analysis of low-grade meta- 
morphic assemblages is presented along 
with experimental data on the join glauco- 
phane + quartz and high-pressure recon- 
naissance on the glaucophane composition. 

Riebeckite-arfvedsonite. Investigation of 
the composition Na 2 ■ 3FeO ■ Fe 2 Oa " 8S1O2 


+ excess water is being carried out under 
controlled partial oxygen pressures using 
the buffer technique of Eugster (Year 
Book 55). At high Po 2 and low tempera- 
ture, blue riebeckite, °Na 2 Fe 3 ++ Fe 2 +++ Si8- 
022(OH)2, has been obtained, but at lower 

ing Po 2 lowers the high-temperature limit 
of riebeckite-arfvedsonite, it elevates the 
upper stability limit of acmite. This con- 
dition results from the fact that the acmite- 
bearing assemblage is the smallest-volume 
(most oxidized) assemblage; consequently 

Temperature, °C 

Fig. 24. Preliminary Po 2 -T diagram for the composition Na 2 - 3FeO - Fe 2 3 - 8Si0 2 with excess 
water at 2000 bars vapor pressure. Abbreviations: Ac, acmite; Fa, fayalite; H, hematite; L, liquid; 
M, magnetite; Q, quartz; and V, vapor. Buffer curves are light lines except where they coincide 
with field boundaries. Buffers are: (1) hematite-magnetite, (2) magnetite + silica— fayalite, (3) 
magnetite-wiistite, (4) magnetite-iron, (5) wiistite-iron. 

partial oxygen pressure and higher tem- 
perature, the assemblage quartz 4- green 
amphibole occurs (see fig. 24). From opti- 
cal properties as well as chemical con- 
siderations, it is thought that the green 
amphibole represents solid solution be- 
tween riebeckite and arfvedsonite, Na 3 - 
Fe 4 ++ Fe +++ Si 8 22 (OH)2. 
Figure 24 shows that, although increas- 

increased partial oxygen pressure favors 
expansion of this assemblage at the expense 
of both ferrous-ferric amphibole and the 
ferrous-rich melt. The increased stability 
of quartz at higher Po 2 simply indicates 
that the partial molar volume and en- 
tropy of Si0 2 in the melt are greater than 
as a solid phase. 
P-T relations found in this system make 



it possible to explain certain natural oc- 
currences. For instance, rocks of the 
Quincy, Mass., granite series generally con- 
tain an arfvedsonitic amphibole in contrast 
to late-stage (lower-temperature) pegma- 
tites which carry riebeckite. During crys- 
tallization of the magma, ferrous iron 
(assumed relatively abundant with respect 
to water) probably maintained Fo 2 at a 
moderately low value; in the highly aque- 

preponderance of calciferous amphiboles 
in igneous and metamorphic rocks, bulk 
compositions high in soda and lime-de- 
ficient relative to alumina are necessary for 
the production of glaucophane over a wide 
P-T range. Because of the scarcity of such 
chemical environments, glaucophane ought 
to be very rare. 

Nevertheless, this mineral is characteris- 
tically present in highly deformed portions 

Oxide ratios 

+ Average tholeiitic basalt, 
(Nockolds, 1954) 

• Glaucophane schists 

v Epidote amphibolites 

o Greenschists 


MgO + FeO 
Chemical variation of ten glaucophane schists, greenschists, and epidote amphibolites of 

Fig. 25 
roughly basaltic composition. Oxide ratios are given in mole per cent. 

ous pegmatite stage the predominance of 
H 2 over iron oxide would tend to elevate 
partial oxygen pressure. This relative dif- 
ference in oxidation level is also reflected 
in the ores, magnetite being most common 
in the Quincy granite members (Emerson, 
1917; LaForge, 1932) and hematite in the 
pegmatites (Warren and Palache, 1911). 

The glaucophane schist fades. Experi- 
mentally determined phase relations for 
pure glaucophane as well as for glauco- 
phane + quartz (see following section) in- 
dicate that this amphibole has a wide range 
of thermal stability. To judge from the 

of a number of geosynclines where it has 
developed in rocks of a variety of compo- 
sitions. As an example, figures 25 and 26 
show that many glaucophane schists ex- 
hibit the same compositional range as 
greenschists and epidote amphibolites of 
roughly basaltic composition. Provided 
they represent equilibrium assemblages, 
such greenschists, epidote amphibolites, 
and glaucophane schists must have formed 
under different physical conditions, and 
therefore represent different metamorphic 
The important minerals of most of these 


low-grade rocks include actinolite, albite, 
chlorite, epidote, garnet, glaucophane, iron 
ores, jadeitic pyroxene, lawsonite, musco- 
vite or paragonite, pumpellyite, quartz, 
and stilpnomelane. These phases may be 
described in terms of eight components, 
Si0 2 , Al 2 O a , Fe 2 3 , FeO, MgO, CaO, 
Na 2 0, and H 2 0. Ti0 2 , MnO, K 2 0, and 
C0 2 play relatively minor roles in many of 
these rocks and are therefore neglected. As 
a first approximation, ferric oxide may be 
assumed to be crystallized as ores. FeO 

1 1 



• .— •_ • 


Na 2 

' l 

xv wvv w y SiOj 


WWW w 



Oxides in weight per cent 

MgO). Small tetrahedra represent seven 
phase assemblages, four critical minerals 
with the additional possible phases iron ore, 
quartz, and vapor. For simplicity, stilp- 
nomelane— chlorite— garnet and lawsonite- 
pumpellyite-epidote have been considered 
as two "isochemical" groups. Because their 
compositions are not identical, garnet plots 
at a slightly different location from chlorite 
and stilpnomelane; similarly, lawsonite 
does not plot at exactly the same point as 
pumpellyite and epidote. Thus two or 

1 1 











•• • •> • m • 


V X&& V S7 

v FeO 




v w ww v 

Fe 2 3 

) CD O O O OO 

1 1 

O o 


+ Average tholeiitic basalt, (Nockolds.1954) 
• Glaucophane schists 
v Epidote amphibolifes 
o Greenschists 

Fig. 26. Chemical variation of ten glaucophane schists, greenschists, and epidote amphibolites of 
roughly basaltic composition. 

and MgO are treated as one component, 
FmO, assuming that Fe ++ ^->Mg solid solu- 
tion exists among minerals for the com- 
positional variation attained by these rocks. 
Where this assumption is invalid (e.g., 
garnet-bearing rocks formed at low pres- 
sure), an additional FmO-bearing phase is 
to be expected. Quartz is considered a 
ubiquitous phase; similarly water vapor is 
presumed as always present. 

Making these approximations, low-grade 
metamorphic assemblages can be repre- 
sented in a four-component system. In 
figure 27 phases are plotted in the tetra- 
hedron Al 2 3 -Na 2 0-CaO-FmO (FeO 4- 

three members of each group may stably 
coexist under a restricted range of physical 
and chemical conditions, but the over-all 
P-T fields of stability of individual mem- 
bers of each group are certainly dissimilar. 
Solid solution between glaucophane and 
actinolite has been neglected; analyses of 
natural amphiboles (Sundius, 1946) sug- 
gest only limited solid solution. 

In figure 27a through d, small tetrahedra 
shown with heavy lines enclose the average 
tholeiitic basalt (Nockolds, 1954), and in- 
dicate mineral assemblages obtained in 
rocks of this general composition. Addi- 
tional tetrahedra (giving assemblages stable 


for various other compositions) are indi- 
cated by lighter lines. 

Figure 27a illustrates possible assem- 
blages (as interpreted from field and petro- 
graphic descriptions and from experimen- 
tal studies) stable under conditions of the 
greenschist and epidote amphibolite facies. 
The graphical analysis shows that in nor- 
mal low-grade metamorphism albite is 

tion of sodic amphibole in the chemically 
unusual rocks may be the result of favor- 
able physical as well as chemical con- 

The stability of glaucophane in rocks of 
a wider variety of compositions results 
from join shifts, promoting the formation 
of alkali amphiboles at the expense of 
albite + ferromagnesian minerals. Neglect- 


Fig. 27. Graphical analysis of low-grade metamorphic assemblages. Reactions a-^b — » d and 
a — > c — > d are favored by increasing pressure or decreasing temperature. FmO = FeO + MgO. 
Ab, albite; Ac, actinolite; CI, clinochlore; Ep, epidote; Jd, jadeite; Ga, garnet; Gl, glaucophane; 
Lw, lawsonite; Pa, paragonite; Pm, pumpellyite; Q, quartz; St, stilpnomelane; V, vapor. *, aver- 
age tholeiite basalt of Nockolds; oxides in mole per cent. 

stable in rocks of a wide chemical range, 
but that glaucophane (which cannot co- 
exist with either calcium-aluminum sili- 
cate or white mica) is confined to bulk 
compositions deficient in CaO and rich 
in Na a O (and FeO + MgO) relative to 
A1 2 3 . Examples of such chemically un- 
usual glaucophane schists have been de- 
scribed from Corsica and Switzerland. 
(However, these rocks are associated with 
glaucophane schists and greenschists of 
basaltic composition, so that the crystalliza- 

ing reactions among the "isochemical" 
series (e.g., pumpellyite — » epidote + va- 
por, chlorite — » garnet + vapor) , only two 
major join shifts are possible in the de- 
scribed system: (1) albite + ferromagnesian 
silicate (such as chlorite) —» glaucophane + 
white mica, and (2) albite + actinolite + 
ferromagnesian silicate — > glaucophane + 
calcium-aluminum silicate (e.g., lawson- 
ite). Both reactions involve volume reduc- 
tions and are therefore favored by high 
lithostatic and/or vapor pressure (or low 


temperature). The magnitude of AF de- 
pends on which ferromagnesian and cal- 
cium-aluminum silicates take part in the 
reactions (which in turn largely depends 
on the temperature) but negligible amounts 
of vapor are produced on one side or 
the other of the equations. Reactions 1 and 
2 are univariant P-T curves, and because 
their relative positions are unknown, their 
effects on assemblages of the system are 
shown as alternative steps. In the upper 
path, a-b-d, the assemblage glaucophane + 
white mica is presumed stable at lower 
pressure than the assemblage glaucophane 
+ calcium-aluminum silicate; the lower 
path a-c-d is a possible alternative. Figure 
7ld shows phase relations occurring at 
pressures in excess of both univariant 
curves; here albite is restricted to bulk com- 
positions high in soda content, although 
glaucophane is stable over a wide range 
of chemical conditions. A final high-pres- 
sure reaction is indicated in figure 27d: 
albite — > jadeite + quartz. This reaction 
must occur at higher pressures for any tem- 
perature and bulk composition than reac- 
tions 1 and 2 because diverse natural assem- 
blages contain both glaucophane and albite, 
as illustrated in figure 27£, c, and d. 

It may be concluded that although glau- 
cophane is stable under P-T conditions of 
the greenschist and epidote-amphibolite 
facies, its coexistence with a calcium-alu- 
minum silicate and/or sodic mica and/or 
the presence of the assemblage jadeite + 
quartz results from relatively high pres- 
sures (or low temperatures) characteristic 
of the glaucophane schist facies, and it is 
on the basis of these mineral associations 
that the facies should be defined. The rare 
occurrence of glaucophane in rocks of 
greenschist and epidote amphibolite facies 
is a manifestation of hypersodic and sub- 
calcic bulk compositions; in rocks of more 
normal chemistry, glaucophane results 
either from pressures in excess of those 
commonly attained during low-grade re- 
gional metamorphism or at lower tem- 

The join glaucophane + quartz. Quartz 

occurs in many glaucophane schists and 
must reduce the temperature range of sta- 
bility of the amphibole. Preliminary data 
indicate that excess silica lowers the high- 
temperature limit of glaucophane 10° to 
15° C. Therefore, other factors being 
equal, the physical conditions under which 
glaucophane is stable in a given rock are 
practically independent of its degree of 
Si0 2 saturation. 

The bulk composition glaucophane + 2 
quartz is equivalent to that of talc + 2 al- 
bite. For this composition, glaucophane is 
stable at 2000 bars vapor pressure from 
below 700° C to above the breakdown 
temperature of talc (825° C), indicating 
that the assemblage talc + albite has no 
field of stability under these conditions. 
Thermodynamic data suggest that glau- 
cophane + quartz constitute the low-tem- 
perature, high-pressure assemblage, and so 
presumably occurrences of pure talc + al- 
bite, as in South Australia (Stillwell and 
Edwards, 1951), are either metastable or 
formed at high temperature and low 

High-pressure investigation of glauco- 
phane. Reconnaissance of the glaucophane 
composition at 600° to 800° C and 20 to 30 
kilobars suggests that two polymorphs of 
glaucophane may exist. Unit cell dimen- 
sions of the high-pressure amphibole agree 
well with those of natural glaucophane 
near the °Na2Mg 3 Al2Si 8 022(OH)2 end 
member, but are appreciably smaller than 
those of glaucophane hydrothermally syn- 
thesized at 200 to 2000 bars vapor pressure 
(see table 9). The two amphiboles have 
identical optical properties. Glaucophane 
grown at 700° C and 20 kilobars has been 
rerun at 800° C and 1000 bars Pvapor, where 
it recrystallized to the form with the large 
unit cell. Work is now in progress to 
delineate the P-T fields of the two forms. 


Members of the biotite group of minerals 
are among the more common ferromag- 
nesian constituents of igneous and meta- 
morphic rocks, and a satisfactory under- 


standing of the petrogenesis of such rocks 
necessarily involves knowledge of the 
phase equilibria of their component 

Compositional variations in the biotite 
group are marked by extensive replace- 
ments in a variety of lattice positions. Ex- 
perimental studies must start with the 
simpler end members and gradually work 
toward the complex solid solutions charac- 
terizing the natural minerals. Yoder and 
Eugster (1954) have determined the upper 
stability curve of the magnesian end mem- 
ber, phlogopite, and Eugster (Year Book 
56) has determined the phase relations for 
the ferrous iron end member, annite. 

Studies conducted during the past year 
have established the phase relations of 

TABLE 9. Unit Cell Dimensions of Synthetic 

800° C, 1000 bars 

800° C, 20 kilobars 

a = 9.99 A 

a - 9.83 A 

b— 17.92 A 

£=17.71 A 



3 = 71.6° 

3 = 72.3° 

Cell volume 

Cell volume 

= 895 A 3 

= 875 A 3 

biotites on the join phlogopite-annite, and 
the phase relations of "ferri-annite," a 
biotite with complete substitution of Al +3 
by Fe +3 . 

Biotites on the Join Phlogopite (KMgs- 

AlSi 3 O 10 {OH) 2 )- Annite (KFe 3 AlSi s Oio- 

(OH) s ) 

D. R. Wones and H. P. Eugster 

The general phase relations of the bio- 
tites on this join and their graphical rep- 
resentation were presented in last year's 
report. Phase equilibria involving these 
biotites require, in addition to the normal 
parameters of vapor pressure and tempera- 
ture, the parameter partial pressure of 
oxygen. These biotites may coexist with 
the following assemblages (in order of 
decreasing partial oxygen pressure and in- 
creasing temperature) : (1) sanidine + hem- 

atite + vapor, (2) sanidine 4- magnetite 4- 
vapor, (3) leucite 4- magnetite 4- olivine 4- 
vapor, and (4) leucite 4-kalsilite + olivine + 
vapor. Because of its greater interest in 
connection with the majority of natural 
biotite occurrences, the major emphasis has 
been on the phase relations involving bio- 
tite coexisting with sanidine and magnetite. 

In figure 28, annite content is plotted as 
a function of temperature at partial pres- 
sures of oxygen equivalent to buffering 
systems of magnetite-hematite (fig. 28«), 
nickel-nickel oxide (fig. 2Sb) and fayalite- 
silica-magnetite (fig. 28<r) at a total water 
vapor pressure of 2046 bars, and represents 
the composition of a biotite coexisting with 
sanidine, magnetite, and water vapor at 
the stated conditions. 

The compositions of the biotites were 
determined by both optical and X-ray tech- 
niques. The assemblages were formed 
from two or more starting materials: bio- 
tite; a mixture of biotite, sanidine, and 
magnetite; and a mixture of K2O, Si0 2 , 
AI2O3, MgO, and Fe yielding the desired 
total composition. 

Figure 29 summarizes the data of figure 

28 in a series of curves representing the 
partial pressures of oxygen of biotite-sani- 
dine-magnetite assemblages as a function 
of temperature at a constant water vapor 
pressure of 2046 bars. The curves in figure 

29 represent the temperatures and partial 
pressures of oxygen at which the respective 
biotites are stable with sanidine and mag- 
netite. The relations are limited to the 
regions in which magnetite is the stable 
iron oxide, as no ferrous-bearing biotite has 
been proved stable in the hematite field, 
and the higher-temperature portions of the 
diagram were only partly investigated. 

The phase relations may be clarified by 
examining the triangular diagrams in fig- 
ure 30. In the portions of the system en- 
countered the components necessary to 
describe all phases present are: K2O' 
Al 2 Cv6Si0 2 , MgO, Fe, 2 , and H 2 . By 
considering H 2 as always present in a 
ubiquitous vapor phase (water vapor), the 
system may be expressed by a tetrahedron 



1000 - 

Hematite -Magnetite Buffer 

p vapor = 2046 bars 
Biotite + Vapor 

Biotite + K-Feldspor + Hematite 
Biotite + K-Feldspor + Hematite + 

Magnetite + Vapor 
Composition of Biotite in Equilibrium 
with K-Feldspar + Magnetite + 
Hematite + Vapor 

Biotite + K-Feldspar + Magnetite + Hematite + Vapor 






40 50 60 70 80 

Weight per cent Annite 

Fig. 28a. Temperature-composition relations of biotites on the join phlogopite-annite at a constant 
water vapor pressure of 2046 bars, at partial oxygen pressure equivalent to hematite-magnetite 

~\ r 

Nickel Oxide -Nickel Buffer 


Pvapor -- 2046 bars 




h Biotite + Vapor 


CD Biotite + K-Feldspar + Magnetite + Vopor 




O Composition of Biotite in Equilibrium _ 
with K-Feldspar + Magnetite ♦ Vapor 




Biotite + K-Feldspar ♦ Magnetite + 




Biotite + Vapor 










CD - 




1 1 

1 1 

1 1 1 



40 50 60 

Weight per cent Annite 




Fig. 28r>. 
water vapor 

Temperature-composition relations of biotites on the join phlogopite-annite at a constant 
pressure of 2046 bars, at partial oxygen pressure equivalent to nickel-nickel oxide 



800 - 


600 - 


40 50 60 

Weight per cent Annite 

Fig. 28c Temperature-composition relations of biotites on the join phlogopite-annite at a constant 
water vapor pressure of 2046 bars, at partial oxygen pressure equivalent to quartz-magnetite- 
fayalite assemblage. 

I l l 

1 1 

1 1 

1 1 i i i i 



Vy2orJ s 
/ \ ° 

A \ 




/ \ \ 

-Feldspar i . J 
Magnetite \ 1 S& 
\Biotite i /\«,' 

\ + \ // 
\Vopor y/^S* 





[/ - 







/6 0%\ 










1 / 1 1 1 

i 1 

/ 1 1 

1 1 1 l 1 1 l 


300 400 500 600 700 800 900 1000 1100 

Temperature, °C 
Fig. 29. Composition of biotite coexisting with K feldspar and magnetite as a function of tem- 
perature and partial pressure of oxygen at a constant water vapor pressure of 2046 bars. 



having the corners K 2 • Al2CV6Si02, 
MgO, Fe, and 2 . By projection from the 
oxygen corner we may visualize the system 
in terms of these components, K 2 0* Al 2 3 " 
6Si0 2 , MgO, and "FeO." Figure 30 shows 
the variations in the phase relationships 
resulting from changes in the partial pres- 
sure of oxygen at a constant temperature 
of 700° C and a constant water vapor 

of oxygen to that of the reaction nickel + 
oxygen = nickel oxide yields the phase as- 
semblages of figure 30£. Here we see that 
sanidine may coexist with magnetite but 
remain incompatible with magnesium-iron 
oxides. The magnesioferrite content of the 
magnetite coexisting with sanidine under 
these conditions is less than 20 per cent 
by weight. 


2 W 3 


K 2 0AI 2 3 -6Si0 2 


700 °C 

2046 bars Feldspar + 

Pq ? =IO" 205 bars Biotite + Magnetite 

MgO "FeO 





Fig. 30. Effect of increasing partial pressure of oxygen at 700° C and 2046 bars water vapor pres- 
sure in the system KoO'AloOg^SiOg-MgO-Fe-Oa-H.,. Oxygen pressures are: (a) 10~ 20 - 5 bar, 
(b) 10- 16 - 8 bar, (c) 10" 11 - 3 "bar, (d) 10" - 9 bar. 

pressure of 2046 bars. In figure 30a, the 
partial pressure of oxygen is that of the 
equilibrium reaction wiistite + oxygen = 
magnetite. Consequently, both wiistite and 
magnetite plot at "FeO." Under these 
conditions sanidine reacts with water va- 
por and any of the iron magnesium oxides 
to form biotite. Maintaining temperature 
and total water vapor pressure constant 
but increasing the partial vapor pressure 

The composition of the biotite coexisting 
with sanidine and magnetite is unique for 
specific values of temperature, total pres- 
sure, and partial pressure of oxygen. For 
a fixed total water vapor pressure of 2046 
bars, figure 29 represents the specific tem- 
perature and partial pressure of oxygen at 
which biotite of a given composition may 
coexist stably with sanidine, magnetite, and 
water vapor. 


Figure 30c represents the system at a 
partial pressure coincident with the reac- 
tion magnetite + oxygen = hematite. Con- 
sequently both hematite and magnetite 
plot on the '"FeO" corner. 

Figure 3(W represents a partial pressure 
of oxygen equivalent to the reaction cop- 
per + oxygen = cuprite. Hematite is the 
only stable iron oxide at these conditions. 
No ferrous biotite has been proved stable 
in the hematite region, and those with an 
annite content of 10 per cent or more have 
reacted to form a more phlogopitic biotite 
+ sanidine + hematite. As no magnesio- 
ferrite was formed under these conditions 
it was assumed that the magnesioferrite 
contents of the magnetites encountered in 
other parts of the system are very low. 

The few occurrences of coexisting fer- 
rous biotite and hematite reported have 
contained either potash feldspar, magnet- 
ite, or ilmenite. The presence of potash 
feldspar could indicate the assemblage bio- 
tite + feldspar + hematite, which is coinci- 
dent with the assemblage hematite + mag- 
netite in terms of partial pressure of oxy- 
gen. The best available evidence indicates 
that ilmenite-hematite assemblages repre- 
sent a slightly lower partial pressure of 
oxygen than hematite-magnetite assem- 
blages at the same temperature, so that 
natural occurrences of ferrous biotite, hem- 
atite, and ilmenite are completely con- 
sistent with the present experimental 

Geologic applications. The most impor- 
tant geologic implication of this study is 
that, for a given temperature, water vapor 
pressure, and partial pressure of oxygen, 
there is a unique value of the Fe/Mg ratio 
of a biotite coexisting with K feldspar and 
magnetite. Conversely, for a given tem- 
perature and water vapor pressure, a vapor 
coexisting with biotite, K feldspar, and 
magnetite has a unique H2/H2O ratio. 
Consequently, if rocks exposed to a freely 
circulating aqueous vapor phase contain 
the assemblage biotite-K feldspar-magnet- 
ite, the Fe/Mg ratio of biotite should tend 
to a uniform value throughout. 

This result may be of interest in connec- 
tion with both the development of contact 
aureoles about intrusive igneous rocks and 
the origin of granitic gneisses. The bulk 
composition of a rock will usually define 
the partial pressure of oxygen of that rock. 
Consequently in metamorphic terrains 
each metasedimentary unit should have a 
characteristic partial pressure of oxygen 
and a characteristic mineral assemblage. In 
the absence of a circulating aqueous va- 
por phase, each unit should maintain its 
characteristic assemblages through several 
grades of metamorphism. If, however, an 
igneous intrusion permeates the host rock 
with a vapor phase, this should tend to 
bring all the units to the same partial pres- 
sure of oxygen. 

It is frequently suggested that circulating 
fluids may metasomatize pre-existing sedi- 
ments to form rocks of gneissic composi- 
tions. If such a process has indeed taken 
place, all the units within a metasomatized 
terrain ought to consist of mineral assem- 
blages representing equivalent partial pres- 
sures of oxygen. Similarly, a magmatic 
body crystallizing from a water-rich melt 
should contain mineral assemblages of 
equivalent partial pressures of oxygen. 

This study has been limited to bulk com- 
positions like those of biotite. The addi- 
tion of alumina and silica to these compo- 
sitions should approximate some of the 
more important systems involving biotites 
in natural environments. The reaction bio- 
tite + quartz — K feldspar + pyroxene + va- 
por, for instance, is a critical one in the 
charnockitic rocks. 

Eugster (Year Book 56) has shown that 
the addition of silica drastically reduces the 
stability of annite (the reaction annite — \al- 
silite + leucite + fay alite + vapor is changed 
to the reaction annite ■ + quartz — K feldspar 
+ fayalite + vapor). Preliminary work in- 
dicates that, at 2046 bars water vapor pres- 
sure, phlogopite reacts with quartz at tem- 
peratures at least 300° C below the stability 
limit of pure phlogopite. Here the reaction 
is changed from phlogopite=\alsilite + 



leucite + forsterite + vapor to ph logopite + 
quartz + vapor— pyroxene + melt. 

The reactions of aluminosilicates (kao- 
lin, pyrophyllite, andalusite, kyanite, silli- 
manite) and silica with biotite are of much 
importance in the metamorphism of the 
pelitic schists. With regard to oxidation- 
reduction, some of these complicated natu- 
ral reactions may be analogous to the 
simpler reactions studied in the laboratory. 
The reaction biotite + sillimanite + oxygen 
— muscovite-\- magnetite + quartz, for in- 
stance, may be considered analogous with 
the reaction biotite + oxygen = magnetite + 
K feldspar + vapor. Consequently, in the 
more complex natural assemblages, in 
which biotites coexist with muscovite, mag- 
netite, quartz, and aluminosilicates, the 
Fe/Mg ratios of the biotites should indi- 
cate whether or not there was a freely 
circulating aqueous vapor phase. 

In metamorphic terrains which contain 
stratigraphic units of varying oxygen con- 
tent, mineral assemblages involving biotite 
should vary from assemblages representing 
reducing conditions (iron-rich biotite- 
graphite-magnetite) to those representing 
oxidizing conditions (magnesium-rich bio- 
tite-magnetite-hematite) . Chinner (this 
report) has indeed observed such changes 
in a suite of rocks from Glen Clova, Scot- 
land, indicating that the behavior of the 
biotites within an aluminum-rich system is 
quite similar to that observed in the present 
synthetic studies. 

The continuous "reaction series" of bio- 
tite with K feldspar and magnetite (or 
with muscovite, Al silicate, and magnet- 
ite) may help explain the discordant ages of 
biotites and other minerals from gneissic 
rocks, and the discrepancies between bio- 
tite ages from schists and the muscovite 
ages of pegmatites intruding the schists 
(Tilton et al., this report). Very slight 
changes in temperature or partial pressure 
of oxygen will promote recrystallization of 
biotite assemblages, whereas muscovite- 
feldspar assemblages are stable over rela- 
tively larger variations in temperature and 

D. R. Wones 

The stability relations of "ferri-annite," 
K 2 • 6FeO • Fe 2 3 ■ 6Si0 2 ■ 2H 2 0, have 
been established for 1023- and 2046-bar iso- 
baric sections. The diagrams representing 
these relations are given as figures 31a and 
3>\b, which plot the field boundaries as 
functions of temperature and partial pres- 
sure of oxygen. The phase assemblages 
coexisting with water vapor are (1) "ferri- 
annite," (2) iron sanidine + hematite, (3) 
iron sanidine + magnetite, (4) melt + mag- 
netite, (5) wiistite + fayalite + melt, and 
(6) iron + magnetite + melt. These dia- 
grams show some similarity to those for 
annite in the low-temperature part, but the 
incongruent melting of the potash-iron- 
silica phases at temperatures distinctly 
lower than their aluminum analogues pro- 
duces distinct changes in the high-tem- 
perature part of the diagram. 

A comparison of figure 31£ and figure 
31c shows that the maximum temperature 
at which "ferri-annite" is stable is within 
10° C of that at which annite is stable. 
"Ferri-annite" is stable with hematite and 
magnetite at a higher temperature than 
annite, and in the magnetite region "ferri- 
annite" exhibits higher stability than ann- 
ite in terms of partial pressure of oxygen. 

Biotites containing "ferri-annite" as a 
component have been reported from peg- 
matites, in which the partial pressures of 
oxygen are generally presumed to be 
higher than in other igneous rocks. Al- 
though biotites rich in "ferri-annite" might 
be expected to form elsewhere, as, for 
instance, in iron formations or ferruginous 
quartzites containing potash but poor in 
alumina, no such occurrences have yet been 

H. S. Yoder, Jr., and W. Schreyer 

An extensive field of mullite (3AI2GV 
2Si0 2 ) appears in phase diagrams of an- 
hydrous systems containing AI2O3 and 
SiO-. In nature, mullite is a rare mineral 


occurring in shallow, high-temperature 
contact rocks such as porcellanites and 
buchites. On the other hand, sillimanite 
(Al20 3 'Si02) is a very common mineral 
in high-grade metamorphic rocks. Ken- 


clO" lu - 

-20 _ 

Water vopor pressure: 
1023 bars 

1 Melt + 

1 Hern^""'^ 

1 J?r 

FeSan + Hem / y 

Melt + Mag 


=Melt ♦^2^^- 

-wus^>; Me V 

- /- KFe 3 FeSi 3°in (0H) ?" 

rr/Foyy/ - 

/= P 




, , / . 




600 700 

Temperature. °C 


addition, recent studies in K2O-AI2O3- 
Si0 2 -H 2 (Yoder and Eugster, 1955) and 
MgO-Al 2 03-Si02-H 2 (Schreyer and 
Yoder, this report) suggest that the liq- 
uidus field of mullite is greatly reduced 



600 700 800 




600 700 

Temperature, °C 

Fig. 31. Phase relations of "ferri-annite," plotted as a function of partial pressure of oxygen and 
temperature: (a) "ferri-annite" at 1023 bars water vapor pressure, (i>) "ferri-annite" at 2046 bars, 
(c) annite at 2046 bars. 

nedy (personal communication, 1956) 
indicated in preliminary data that mull- 
ite gradually transforms into sillimanite 
( + corundum) above a curve which passes 
approximately through the points 3000 
bars, 600° C, and 7000 bars, 900° C. In 

with increasing water pressure. New data 
on the behavior of synthetic mullite at 
10,000 bars water pressure demonstrate that 
mullite is not stable at this pressure. As 
low as 675° C mullite breaks down readily 
into sillimanite + corundum. The latter 


phases are stable in the presence of water 
up to about 1050° C, where melting takes 
place, yielding corundum + liquid + vapor. 
It is apparent that the incongruent melting 
of mullite to corundum + liquid which 
takes place at 1810° C at 1 atm (Bowen 
and Greig, 1924) is drastically lowered 
with increasing water pressure. Extending 
Kennedy's preliminary data, the incongru- 
ent melting curve of mullite presumably 
terminates at about 9000 bars water pres- 
sure where sillimanite + corundum + vapor 
appear as the stable assemblage on the 
solidus. Above this water pressure silli- 
manite in turn melts incongruently to co- 
rundum + liquid. 

A. C. Turnoc\ 

Minerals of the spinel group are found 
in nearly every type of rock, usually as a 
minor mineral, sometimes in major 
amount. As many as three spinel phases 
may coexist in a rock, but one is the usual 
number. The general formula, XsYieO.^ 
(with X=Fe, Mg, Ni, Mn, Zn; Y=A1, Fe, 
Cr, Ti), indicates the magnificent possibili- 
ties for solid solution. The actual composi- 
tion of any naturally occurring spinel min- 
erals must reflect a limit of solid solution 
imposed by the chemical and physical con- 
ditions of the environment of formation, 
but these limits may be very difficult to fix. 

Magnetite, Fe 3 4 , the most ubiquitous 
member of the spinel group, occurs in both 
igneous and metamorphic rocks. Hercy- 
nite, FeALOi, occurs in metamorphosed 
aluminous sediments, spinel-emery de- 
posits, and basic igneous rocks, and is on 
the whole uncommon. The amount of 
solid solution between these two end mem- 
bers is limited by a solvus (or two-phase 
region) which has been determined to 
close in the geologically interesting tem- 
perature range 500° to 900° C. Subsolidus 
relationships in the system Fe-Al-O show 
how the magnetite-hercynite series is re- 
lated by oxidation and reduction reactions 
to hematite, corundum, wiistite, and iron. 
Using the hydrothermal technique, with 

Ptotai = PH 2 o + PH 2 , we have worked up to 
4000 bars, in the temperature range 500° 
to 900° C, using a solid buffer to control 
thePo 2 (Eugster, 1957). 

The effect of composition on the size of 
the edge of the unit cell in the magnetite- 
hercynite series is shown in figure 32. The 
close agreement with the published results 
of previous workers indicates that tem- 
perature and pressure have very little effect 
on cell size, except in the cation-deficient 
region which lies at high oxygen pressure 
and temperature greater than 900° C. 

The effect of temperature on the solid 
solution between magnetite and hercynite 
is shown in figure 33. Above 858° ±8° C 
complete solid solution exists. Below this 
temperature the two-phase region of ex- 
solution widens with decreasing tempera- 
ture. The reactions are very sluggish, so 
that equilibrium cannot be sharply de- 
fined, even at 2000 bars pressure. The effect 
of pressure on this solvus is very small, and 
cannot be detected in the range 1 to 4000 

The effects of changes in the partial pres- 
sure of oxygen are oxidation and reduction 
reactions that change the species and bulk 
composition of the solid phases. These re- 
lationships can best be shown by a three- 
dimensional sketch of the system Fe-Al-O, 
with compositions and Po 2 as coordinates 
(fig. 34) . The basal triangle gives the bulk 
composition of the solids. The complex 
surface, ruled horizontally, with slopes, 
cliffs, and plateaus, is the surface repre- 
senting solid assemblages that are in equi- 
librium with vapor, for that partial pres- 
sure of oxygen given by the vertical 
coordinate. This diagram shows the relative 
stability ranges of the minerals, in terms 
of Po 2 > and illustrates the fact that large 
changes in the Po 2 may cause radical 
changes in the mineral assemblage. Her- 
cynite has only a small stability range, and 
at a lower Po 2 than that of magnetite. At a 
Po 2 slightly higher than the maximum for 
pure ferrous hercynite, the stable assem- 
blage will be ferrian herycnite + corundum. 
The tie lines between these two minerals 



1 1 

1 1 l 1 1 l 1 




Hoffmann ft Fischer 
1000° 8 1250° 


^ S% ^» 

Atlos 8 Sumida 
^'s^. 800-850° hydrofhermal 




^f^- Turnock a Eugster 







1 t 

1 1 1 1 1 1 1 



Mole per cent Hercynite 


FeAl 2 4 

Fig. 32. Cell edge versus composition for the spinel solid solution series magnetite to hercynite 
(Hoffmann and Fischer, 1956; Atlas and Sumida, 1958). 



O 700 

u 600 




"] I i 1 1 1 r 

l 1 1 1 1 1 1 1 1 r 

►o $► o -j <s 

► ► QO 4 <J 


Spinel solid solution 

t> ««J>003 H 

Magnetite ss + Hercynite ss 

J I I L 

o One phase 
►« Solid solution 
"" Exsolution 

Grown from oxide mixes 
_l I I I I I I I i I i 






Weight per cent Hercynite 


Fig. 33. Effect of temperature in limiting the amount of solid solution in the spinel series, mag- 
netite to hercynite. Isobaric section (P t otai — 2000 bars) . 


form a sloping surface, a relationship in 
which continuous changes of Po 2 cause con- 
tinuous changes in the composition of the 
minerals. A further increase in the Po 2 
leads to the oxidation of the ferrian hercy- 
nite to aluminum magnetite + ferrian co- 
rundum. As long as the three solid phases 
are present, they buffer any changes in Po 2 ; 
this buffer effect (univariant equilibrium) 
is represented on the diagram by a plateau. 

Exsolved hercynite in magnetite has been 
reported from basic igneous rocks (Ram- 
dohr, 1955, pp. 699-700). The solvus can- 
not be used to interpret the thermal history 
without chemical analyses of the phases, 
but it can be inferred that the spinel crys- 
tallized as one phase at a high temperature, 
of the order 700° to 900° C. Exsolution 
would occur at a lower temperature, sug- 
gesting that the rock held a temperature 

Fig. 34. The surface of solid phases in equilibrium with vapor. The system Fe-Al-O, isothermal 
(T «* 700° C), isobaric (P tota i = 2000 bars). A, iron + corundum; B, hercynite + corundum; C, 
ferrian hercynite + corundum; D, hercynite field; E, hercynite ss + magnetite ss; F, magnetite field; 
G, magnetite ss + corundum ss; H, hematite ss + corundum ss. 

The subsolidus relations are summed up 
in figure 35, in two isobaric, isothermal sec- 
tions. Heavy lines, representing the vari- 
ability in composition of the minerals, are 
regions of one solid phase. The tie lines 
connecting equilibrium assemblages in re- 
gions of two solid phases are also contour 
lines of the partial pressure of oxygen, with 
values increasing up to the hematite-co- 
rundum join. 

The phase relations given in figure 35 
can be related to naturally occurring min- 
eral assemblages; examples follow: 

in the range 400° to 600° C for a long 

As an example of the assemblage her- 
cynite + corundum, Tilley (1924) has de- 
scribed the assemblage hercynite + corun- 
dum + ferroan cordierite from the hornfels 
in the thermal aureole at the edge of a 
diorite plug in the Comrie area, Scotland. 
This rock formed in an environment that 
had an unusually low Po 2 , as can be seen 
from figure 34, and which is confirmed by 
analysis showing only a trace of ferric 
iron. If it can be assumed that the original 



sediment contained only ferric iron, which 
is usual for material rich in aluminum and 
iron, then it has been reduced during the 
successive regional and thermal meta- 
morphic cycles. 

The assemblage hercynite + magnetite + 
corundum is represented by a hornfels 
composed of these minerals, with minor 
biotite + K spar, described by Read (1931) 
from the thermal aureole of a minor basic 
intrusive in the Haddo House area, Scot- 
land. Figure 34 indicates that this rock 
formed under a higher Po 2 than the her- 
cynite + corundum rock of the previous ex- 

Wones (Year Book 57, p. 210) has de- 
scribed the assemblage magnetite + corun- 
dum + hematite from the emery ore at 
Chester, Mass. This example must have 
formed in an environment very similar to 
that of the previous example, except for a 
slightly higher Po 2 - 

The assemblage hematite + corundum 
was described by Agrell and Langley 
(1958) in a xenolith in a basic plug, at 
Tievebulliagh, Ireland. A second example 
is the assemblage hematite + corundum 
with minor diaspore + chloritoid from an 
emery deposit in Turkey. In both assem- 


Weight per cent 




FeAI 2 4 




FeAI 2 4 


Fig. 35. Subsolidus relationships in the system FeO-Fe 2 3 -Al,0 3 . Isothermal sections (900° and 
700° C), isobaric (P total « 2000 bars). 

ample. A temperature determination could 
be made from the solvus only if the com- 
position of the two spinels was known. 

As an example of the assemblage mag- 
netite + corundum, a sample of emery ore 
from the deposit at Chester, Mass., was 
found to be composed of corundum + mag- 
netite (nonaluminous), with minor chlo- 
rite. The purity of the magnetite indicates: 
(1) a partial pressure of oxygen only 
slightly lower than that of the hematite- 
magnetite buffer (see fig. 34; the magnet- 
ite-corundum tie line must be at the 
maximum Po 2 of the magnetite stability 
range to give pure magnetite) ; and (2) a 
temperature less than 500° C (the mag- 
netite-corundum tie line only reaches pure 
magnetite below this temperature). 

blages, lateritic sediments have been meta- 
morphosed without reduction. The pres- 
ence of diaspore in the emery implies 
hydration of the alumina in the tempera- 
ture range 300° to 400° C, with consider- 
able water vapor pressure. 


C. R. Robbins? H. S. Yoder, Jr., and 
J. F. Schairer 

Tourmaline is the most abundant and 
geochemically the most important of the 
boron minerals. It is found in a variety of 
igneous, metamorphic, and sedimentary 
rocks of all ages. Tourmaline ranks with 
zircon in abundance and frequency of oc- 

? ' U. S. National Bureau of Standards. 



currence as an accessory mineral in sedi- 
ments. Its reported authigenic formation 
in some sandstones is of considerable in- 
terest. Chemically, tourmaline is a com- 
plex borosilicate of variable composition. 
The variations result from the large num- 
ber of possible ionic substitutions occurring 
in the central "pocket" of a relatively fixed 
framework of oxygen, silicon, aluminum, 
and boron. Preliminary calculations sug- 
gest that tourmalines may be described as 
isomorphous mixtures of several end mem- 
bers. The important iron-free end mem- 
bers are dravite, NaMgsAlsBsSieGv- 
(OH) 4 , and uvite, CalV^AlsBsSieC^T- 
(OH) 4 . 

Previous work on tourmaline synthesis 
has been directed primarily toward obtain- 
ing large crystals suitable for commercial 
use in piezoelectric gauges. Its pressure- 
temperature stability range was not de- 
termined. The scope and objectives of the 
present study are: to determine the pres- 
sure-temperature stability range of the end 
members dravite and uvite; and to deter- 
mine the extent of solid solution between 
these end members. 

Both dravite and uvite have been syn- 
thesized in this laboratory, and some in- 
formation on the stability of dravite has 
been obtained. Synthetic dravite has been 
prepared from glass and powdered mix- 
tures of the requisite composition. Be- 

cause of the volatility of B2O3 from dry 
melts, a mechanical mixture of B2O3 glass 
and a glass of the remaining constituents 
was found to be most suitable. Prisms of 
dravite up to 0.3 mm in length have been 
obtained. Indices of refraction of the pure 
dravite end member are : u = 1.632 and £= 
1.610 ( ± 0.002) . The powder X-ray diffrac- 
tion pattern was indexed on the basis of 
space group Rim (Buerger and Parrish, 
1937), and the unit cell dimensions referred 
to hexagonal axes are: a = 15.93 A, C—7.1S 
A, with c/a = 0.4507. The axial ratio from 
X-ray data agrees well with that from pre- 
vious X-ray and morphological studies. A 
natural tourmaline (Mesa Grande, Cali- 
fornia) near dravite in composition ap- 
peared to melt incongruently at about 700 
C and 2000 bars. 

Crystals of synthetic uvite obtained thus 
far are in the form of spherulites approxi- 
mately 0.1 mm in diameter and are un- 
suitable for index of refraction determina- 
tions. The unit cell dimensions are a — 
15.86 A and c=7.19 A with c/a=0A533. 

A number of natural tourmalines have 
been examined optically and with X rays 
in a search for crystals close to the pre- 
dicted end members. Specimens from 
Dobruwa (or Dobrova), Austria; Gouver- 
neur, New York; DeKalb, New York; and 
Ceylon have been found to be very similar 
to the iron-free synthetic end members. 


During this past year continued sys- 
tematic laboratory studies of the phase 
relations among the common sulfide-type 
minerals and applications of the experi- 
mental data to ore deposits have added 
significantly to our understanding of natu- 
ral mineral assemblages. 

Investigations of the subsolidus relations 
in the Ni-S system have been completed. 
The results, in addition to those already 
obtained in the Fe-S system, have provided 
the basic information for methodical ex- 
ploration in ternary systems involving these 
elements. In the system Fe-Ni-S the phase 

relations between pyrite and vaesite in the 
presence of excess sulfur have been de- 
termined between 300° and 1000° C. This 
information, in addition to the data we 
have acquired already in the Fe-S and 
Ni-S systems, is of great importance to 
the studies now under way on the sulfur- 
deficient regions of the Fe-Ni-S system. 

The upper stability curve of the pyrite- 
arsenopyrite assemblage has been deter- 
mined up to 2000 bars, completing the in- 
vestigation of the Fe-As-S system. A de- 
tailed study of the phase relations in the 
Ni-As-S system is nearly completed. Gers- 


dorffite (NiAsS), the only ternary phase, 
occurs commonly with rammelsbergite (or 
pararammelsbergite) in nickel deposits. 
Extensive solid solutions have been found 
to exist between gersdorffite and the latter 
minerals and in the future may well be 
used in geological thermometry. 

The phase relations determined at 600° 
to 800° C in the synthetic Co-Ni-Fe-As 
system have been applied to the mineral as- 
semblages found in nature. Although gen- 
eral agreement exists between the syn- 
thetic and natural assemblages, some of the 
evidently low-temperature phase relations 
displayed in ores cannot be inferred from 
extrapolation of the synthetic relations ob- 
tained at elevated temperatures. 

The experimentally determined pyrrho- 
tite-pyrite phase relations were applied to 
systematically collected specimens from the 
Coeur d'Alene district, Idaho. Tempera- 
tures of formation of mineral pairs from 
62 localities in the Highland Surprise Mine 
varied between 370° and 490° C and 
agreed, in 14 cases where checks could be 
made, fairly well with temperatures ob- 
tained from coexisting sphalerite. 

Estimates of the temperatures of forma- 
tion of certain sulfide veins by the sphal- 
erite-pyrrhotite method indicated the ex- 
istence of strong temperature gradients 
during formation of the ore. The data ob- 
tained in the Wellington Mine, Colorado, 
permitted a fresh attack on the thermal 
problems associated with this type of ore 
deposition. Although idealization is neces- 
sary for mathematical treatment, and ac- 
curate results cannot be expected, heat con- 
duction calculations, using the estimated 
temperature gradient, suggest that the ore 
solution must have contained 0.1 to 10 per 
cent sulfides. 

Two methods have been explored for 
measurements of vapor pressures over sul- 
fide-type minerals and mineral assem- 
blages. One of these methods, using a 
silica glass U tube containing a eutectic 
mixture of LiCl and KCl which melts at 
350° C, is described below. 

An apparatus for differential thermal 

analysis of sulfide-type minerals has been 
designed to operate at pressures exceeding 
60 bars with temperatures up to 1050° C. 

The study of the system ZnS-H 2 S-H 2 
has shown that the solubility of ZnS in 
the aqueous phase is negligible. However, 
a water-saturated sulfide liquid, immiscible 
with water and of high sulfur content, may 
present a possible medium for ore trans- 

G. Kulleri{d and R. A. Yund 

The compounds in this system are 
Ni 3 S 2 , NirSe, Nii-*S, Ni 3 S 4 , and NiS 2 . Of 
these, heazlewoodite (Ni 3 S 2 ), millerite 
(Nii-zS), polydymite (NisSi), and vaesite 
(NiS 2 ) are well established as minerals. 
Heazlewoodite occurs as a hydrothermal 
mineral with pentlandite, (Fe,Ni)9Ss, in 
serpentinized peridotites. Millerite occurs 
as a primary mineral in many ore deposits 
but is most frequently found as an altera- 
tion product of other nickel minerals. 
Polydymite usually occurs with other sul- 
fides and arsenides such as millerite, gers- 
dorffite, and ullmannite in hydrothermal 
vein deposits. Vaesite often forms as an 
alteration product of pentlandite. It has 
also been reported as a primary mineral 
deposited from relatively low-temperature 
solutions. It commonly contains an ap- 
preciable amount of iron in solid solution 
replacing nickel. 

Heazlewoodite (Ni 3 S 2 ) is readily syn- 
thesized in the dry way in silica tubes by 
mixing nickel with appropriate amounts 
of sulfur and heating the mixtures. Fig- 
ure 36 shows that at 550° ±10° C Ni 3 S 2 
goes through a high-low inversion. Hea- 
zlewoodite, the low-temperature form, has 
hexagonal crystal structure, and shows very 
little tendency to form solid solutions on 
either side of the Ni 3 S 2 composition. Vari- 
ous experiments showed that even at 500° 
C it cannot accommodate as much as 0.5 
weight per cent of either nickel or sulfur. 
The high-temperature phase has a higher 
crystal symmetry than heazlewoodite as 
shown by high-temperature X-ray powder 



patterns. This high form cannot be 
quenched, but its presence is indicated by 
exsolution textures observed in polished 
sections. Its possible existence was sug- 
gested first by Bornemann (1908). This 
phase, in contrast to heazlewoodite, has a 

rather wide field of stability; thus at 600° 
C homogeneous solid solution exists over 
the range from 23.5 to 30.5 weight per cent 
sulfur. The composition of this phase can 
be expressed chemically by the formula 
Ni3±.rS2, but whether such a formula is 

40 50 60 

Weight per cent S 

Fig. 36. Phase relations in the system Ni-S. All phases or phase assemblages are in equilibrium 
with vapor. The condensed phases occurring in the numbered fields of the diagram are as follows: 1, 
L; 2, L + Ni ss; 3, Ni + Ni 3±a ,S 2 ; 4, L + Ni 3±a ,S 2 ; 5, Ni 3 S 2 + Ni 3±a ,S 2 ; 6, Ni + Ni 3 S 2 ; 7, Ni 3 S 2 + 
|3Ni 7 S 6 ; 8, Ni 3 S 2 + otNi 7 S 6 ; 9, Ni 3 S 2 + Ni 3±a ,S 2 ; 10, Ni 3±a ,S 2 + ctNi 7 S 6 ; 11, Ni 3±a ,S 2 + aNiS; 
12, L + aNi^S; 13, ctNi 7 S 6 + ctNiS; 14, (3Ni 7 S 6 + ctNiS; 15, [3Ni 7 S 6 + (3NiS; 16, 0Ni 1HI S + 
Ni 3 S 4 ; 17, Ni 3 S 4 + NiS 2 ; 18, a + pNi^S; 19, aNi^S + Ni 3 S 4 ; 20, aNi la S + NiS 2 ; 21, L + 
aNi^S; 22, L + NiS„; 23, L + NiS 2 ; 24, 2 liquids; 25, NiS 2 + L; 26, NiS 2 + S; 27, L + S. 


correct also from a structural point of view 
is not known. The incongruent melting 
relations of Ni 3 ±a;S2 are shown in figure 36. 
At the incongruent melting point, which 
is situated at 806° ±3° C, the four phases 
NisizSa + NiS + L+F are all stable, ful- 
filling the requirements of invariancy in a 
two-component system. 

A eutectic between Ni and Ni 3 ±ax> 2 , 
earlier reported in the literature to be 
situated at 645° C and 21.5 weight per cent 
sulfur, was found to lie at 637° ±3° C. 

A number of phases such as NieS 5 , 
Ni 7 S 6 , and Ni 9 S 8 are reported in the litera- 
ture to occur in the composition range be- 
tween M3S2 and NiS. Our experiments 
with mixtures of nickel and sulfur in vari- 
ous ratios between 3/2 and 1/1, conducted 
at temperatures from 200° to 1000° C and 
lasting from a few hours to 8 months, pro- 
duced Ni 7 S 6 as the only intermediate phase 
between Ni 3 S 2 and NiS. The Ni 7 Se phase 
shows a small but measurable amount of 
solid solution with Ni 3 S 2 , but no measur- 
able solid solution with NiS. It occurs in 
two crystallographic modifications. The 
low form is stable up to 399° ±2° C, where 
it inverts to a high-temperature form 
which in turn decomposes to Ni 3 ±a;S2 and 
NiS at 573° ±2° C. The crystal structures 
of high and low Ni 7 S6 have as yet not 
been determined. 

NiS and the members of the Nii-^S 
solid solution series are also readily synthe- 
sized from the elements. As seen from 
figure 36, NiS and the Nii-^S mix-crystals 
occur in two crystallographic forms. The 
low-temperature form millerite has a 
rhombohedral crystal structure. Stoichio- 
metric NiS inverts to a high-temperature 
hexagonal NiAs structure at 379° ±3° C, 4 

4 The high-low inversion and melting point of 
stoichiometric NiS were given in earlier Year 
Books as 374° ± 3° C and 800.5° ±1° C, re- 
spectively. These data, which are lower by 5° C 
for NiS inversion and lower by 7° C for NiS 
melting than the present values, were obtained 
using nickel of 99.8 per cent purity containing 
0.1 per cent cobalt. The present data were ob- 
tained by using nickel of 99.99 per cent purity. 
The 0.1 per cent cobalt explains the discrepancies 
in the NiS inversion and melting temperatures. 

whereas Nii-jS with maximum Ni de- 
ficiency inverts at 280° ±5° C. 

The solvus defining the composition of 
Nii-^S when in equilibrium with NiS 2 
was determined at temperatures from 350° 
to 650° C by Arnold and Kullerud (Year 
Book 55, pp. 178-179) . The curve has now 
been determined to 982° ±3° C, which is 
the eutectic temperature between the 
Nix-^S and NiS 2 phases. Nii-j-S in equi- 
librium with N1S2 at 982° ±3° C contains 

61.4 weight per cent Ni, and at 745° ±3° C 
Nii-;rS in equilibrium with NiS 2 contains 

62.5 weight per cent Ni. This solvus curve, 
therefore, is very steep, the Ni deficiency 
increasing by only 1.1 weight per cent 
over almost 250° C. 

The melting relations of the Nii-^S 
solid solution series are also shown in 
figure 36. These relations are very similar 
to those observed earlier for Fei-aS. It is 
noted that only one member, the one con- 
taining 62.3 weight per cent Ni, of the 
solid solution series melts directly to a 
liquid of the same composition as the solid 
phase. This maximum melting point, 
where the solidus and liquidus curves in- 
tersect each other, is at 992° ±3° C. 

The Ni 3 S4 phase has a cubic crystal 
symmetry. It was synthesized by heating 
nickel and sulfur, or NiS and sulfur, in 
stoichiometric proportions at temperatures 
from 240° to 302° C from 10 to 50 days. 
The reaction never was completed, since 
Nii- a S and NiS 2 always occurred with the 
Ni 3 S4 phase. The amount of Ni 3 S4 was 
increased markedly by repeated grindings 
and reheatings. After four regrindings at 
302° C and after 50 days of heating, it was 
found that 80 per cent of the material was 
polydymite (the remaining 20 per cent 
being Nii-^S and NiS 2 ) . Heating of mix- 
tures, as mentioned above, at temperatures 
above 302° C as well as heating of syn- 
thetic Ni 3 S4 at 305° C for 2 weeks pro- 
duced Nii-jS and NiS 2 . It therefore ap- 
pears that Ni 3 S 4 under these conditions is 
stable below 303° ±3° C. At this tempera- 
ture the invariant assemblage Nii_;z;S4- 
NisSi4-NiS 2 4-vapor occurs. 



A eutectic was found to exist between 
Nii-tfS and NiS 2 . The eutectic is situated 
at 982° ±3° C and about 60.2 weight per 
cent Ni. Differential thermal analyses in- 
dicate that NiS 2 melts congruently at 
about 1010° C. The melting relations of 
this compound are also shown in figure 36. 
Vaesite does not grow readily when mix- 
tures of metal and sulfur are heated. It 
forms readily, however, when NiS or 
Nii-zS and sulfur in appropriate amounts 
are used as starting materials. NiS 2 was 
synthesized in this way at temperatures 
ranging from 200° to 1000° C. X-ray 
powder patterns of the quenched products 
always showed the same cubic crystal 

Earlier literature indicates that the 
metal-to-sulfur ratio in vaesite may vary 
considerably. This was investigated by 
tube-in-tube runs (method described in 
previous Year Books) which showed that 
even at 700° C the possible variation is 
within the limits of NiS 2 .oo±o.oo5. Powder 
X-ray diffraction studies of NiS 2 grown at 
700° and 900° C (a) with excess sulfur 
and {b) in equilibrium with Nii-^S gave 
the following results: at 700°, 0=5.6890 
A with excess sulfur and 5.6893 A with 
Nii-tfS; at 900°, a = 5.6882 A with ex- 
cess sulfur and 5.6882 A with Nii-aS. 
These values are in close agreement with 
each other, and all lie well within the lim- 
its of error of the X-ray method. They 
are also in agreement with the a values 
given by Arnold and Kullerud (Year 
Book 55) for NiS 2 grown at temperatures 
from 350° to 854° C and at pressures rang- 
ing from a few millimeters of mercury to 
2000 bars. Even small variations in the 
nickel-to-sulfur ratio of vaesite would prob- 
ably cause a corresponding change in the 
unit cell dimensions. The constancy of 
unit cell dimensions of vaesites, synthe- 
sized over the wide range of conditions 
discussed above, indicates strongly that 
vaesite cannot change its nickel-to-sulfur 
ratio measurably from the NiS 2 composi- 
tion. This result, of course, agrees with 

the findings from tube-in-tube experi- 

Freezing-point determinations of nickel 
and sulfur mixtures containing from 45 to 
less than 20 weight per cent Ni indicate 
the existence of a two-liquid (plus vapor) 
field existing above about 1000° C in this 




L. A. Clar\ and G. Kullerud 

Investigation of the phase relations be- 
tween pyrite (FeS 2 ) and vaesite (NiS 2 ) 
was undertaken as part of an extensive, 
systematic study involving the entire Fe- 
Ni-S system. This particular effort is of 
prime theoretical interest and of consider- 
able economic importance. It involves 
pyrite, the commonest naturally occurring 
sulfide, and the tendency of this mineral to 
take nickel into solid solution. In addition, 
this investigation is the key to the bravoite 
([Fe,Ni]S 2 ) problem, and also yields the 
solid solubility of iron in vaesite. One of 
the potential benefits of determining the 
solvus curve for vaesite is a geological 
thermometer that would be suitable for 
estimating formation temperatures of co- 
existing assemblages of pyrite and vaesite 
(or bravoite). 

In the system Fe-Ni-S, a sulfur-rich 
liquid phase persists down to a tempera- 
ture only slightly lower than the melting 
point of sulfur (114.5° C for pure S). The 
present work is an investigation of the 
solid phases in equilibrium with this liq- 
uid, plus a vapor phase, at temperatures 
between 300° and about 1000° C. 

The study was carried out using sealed, 
evacuated, rigid silica glass tubes; hence 
vapor was present as a phase in all runs. 
If a bulk composition lying on the FeS 2 - 
NiS 2 join is heated, part of the disulfide 
charge decomposes to a monosulfide plus 
a vapor, maintaining a vapor pressure in 
equilibrium with the solid phases at any 
given temperature. To avoid this decom- 


position problem, and to speed reaction 
rates, enough excess sulfur was added so 
that a liquid was present during all runs. 
This liquid phase consists essentially of 
sulfur. Even at temperatures of 800° to 
1000° C it contains only about 0.1 per 
cent iron and nickel. 

Vaesite was reported to be isostructural 
with pyrite by de Jong and Willems 
(1927). Figure 37 shows the variation in 
the cubic cell dimensions of pyrite and of 
vaesite as functions of their nickel and 

in vaesite. In figure 37 it should be noted 
that the cell values for the two solid solu- 
tion series do not coincide with a straight 
line drawn between the values for the pure 
end members. 

Figure 38 shows the phase relations be- 
tween FeS 2 and NiS 2 as determined in 
the presence of sulfur liquid and vapor. 
The plotted compositions on the solidus 
and subsolidus curves have been deter- 
mined using the determinative curves 
shown in figure 37. Solubilities in the two 



~ 5 550 

One phase 

Two phases at 700 °C 

Weight per cent NiS2 

Fig. 37. Variations in the cubic cell values of pyrite and vaesite as functions of their nickel and 
iron contents, respectively. 

iron contents, respectively. The runs used 
to establish the curves were performed at 
700° C with the exception of the two most 
iron-rich vaesites, which were homoge- 
nized at 723.5° C. The runs were heated 
at successively longer intervals until there 
were no further changes in the interplanar 
spacings as determined by X-ray powder 
diffraction techniques. Four individual 
measurements were made from diffractom- 
eter traces on each of two reflections, with 
Lake Toxaway quartz as an internal stand- 
ard, and using CuKa radiation. The 
weighted mean cell values were calculated 
using the (311) and (222) reflections in 
pyrite and the (311) and (320) reflections 

solid solution series reach maxima at 729° 
±3° C. Above this temperature pyrite can 
no longer exist in equilibrium with vaesite 
under these conditions. At 729° ±3° C 
the tie line between pyrite and vaesite is 
replaced by one joining liquid sulfur and 
a homogeneous monosulfide solid solution. 
Hence, at 729° ±3° C the phases pyrite, 
vaesite, (Fe,Ni) i-^S, liquid, and vapor co- 
exist, fulfilling the requirements of invari- 
ancy in a three-component system. 

The phase relations in the sulfur-rich 
portion of the Fe-Ni-S system below 729° 
±3° C are represented in figure 39a. The 
assemblages stable between 729° ±3° C 
and 743° ±3° C are shown in figure 39b. 


Here the pyrite-vaesite tie line has been re- 
placed by the monosulfide-liquid tie line. 
At 743° ±3° C pyrite melts incongru- 
ently, and the phase relations above this 
temperature are represented in figure 39c. 
It must be pointed out that the diagrams 
in figure 39 are semischematic, since each 
represents a range of temperatures and 
the compositions of the ternary monosul- 
fide solid solutions coexisting with pyrite 
and/or vaesite are known only approxi- 
As seen in figure 38 there is no stable 

seem to be closely approaching equilib- 
rium at all temperatures from the solidus 
down to 400° C after 62 days of heating. 
Similar runs at 300° C are far from equilib- 
rium. Unfortunately, on the pyrite side it 
is not clear whether equilibrium has been 
reached at any temperature. Pyrite can be 
readily synthesized from (Fe,Ni)i-a;S solid 
solutions plus sulfur at all temperatures 
down to 300° C. However, in the process 
as much as 8 mole per cent NiS 2 goes into 
metastable solid solution. Upon further 
heating the pyrite gradually exsolves nickel 


°. 700 

(Fe.N0,_ x S 

.Py + lFe.Nii^S ,, 

(Fe,Ni),_ v S + Vs 

p y/ 

Py + Vs 


FeS 2 

NiS 2 

Weight per cent MS2 

Fig. 38. Phase relations between pyrite and vaesite as determined in the presence of a sulfur- 
rich liquid and vapor. Py, pyrite; Vs, vaesite. 

phase (Fe,Ni)S 2 with Fe and Ni in equal 
amounts, which is the composition re- 
ported for the mineral bravoite. Vaesite 
can accommodate a maximum of about 
29.8 mole per cent FeS2 in solid solution. 
It is possible that many of the naturally 
occurring "bravoites" are actually vaesite 
solid solutions or two-phase mixtures of 
nickel-rich pyrite and iron-rich vaesite. 

Time versus composition studies are be- 
ing carried out at a number of different 
temperatures, employing both solution and 
exsolution runs, in order to determine the 
equilibrium positions of the solvus curves. 
The runs giving the vaesite solvus curve 

in the form of iron-saturated vaesite. Ap- 
parently, heating periods considerably in 
excess of 62 days are required to approach 
the equilibrium pyrite solvus curve at all 
temperatures below 700° C. 

The vaesite solvus curve can be used as 
a geological thermometer when pyrite co- 
exists with the vaesite. In the temperature 
range 729° to 500° C the vaesite solvus 
curve has a relatively gentle slope and is 
a sensitive temperature indicator. For in- 
stance, at 600° C the temperatures can be 
determined to approximately ±5° C. Be- 
low 500° C the curve steepens and the 
thermometer is much less sensitive; thus 


(a) T<729 ±3°C 

-\Ni, V S 

(b) T = 729± 3 to 743+ 3°C 

(c) T> 7 43 ±3°C 

Py + (Fe,Ni),_ x S 
+ L +V 


Fig. 39. Phase diagrams depicting the assemblages stable within the specified temperature ranges. 
Po, pyrrhotite; Py, pyrite; Vs, vaesite; L, liquid; V, vapor. 

at 400° C an error of 

:30° C might be 




L. A. Clar\ 

In nature the most common association 
of minerals whose compositions lie in the 
Fe-As-S system is that of pyrite (FeS 2 ) 
and arsenopyrite (FeAsS). The assem- 
blages pyrrhotite (Fei-jS)-pyrite-arseno- 
pyrite, and pyrrhotite-arsenopyrite are 
often seen but are less common. 

The phase relations in the Fe-As-S sys- 

tem were described in last year's report 
(pp. 229-232), where it was shown that the 
pyrrhotite-arsenopyrite assemblage is sta- 
ble up to the melting temperature of ar- 
senopyrite. The invariant, incongruent 
melting temperature of arsenopyrite is 
702° ±3° C. This value was determined 
in sealed, evacuated, silica glass tubes 
where a vapor phase was always present. 

Consideration of the phase relations be- 
tween pyrite and arsenopyrite is of greater 
interest to ore petrologists. This mineral 
pair is unstable in the presence of its satu- 
rated vapor at temperatures above 491° 


±12° C. 5 At 491° ±12° C the five phases 
pyrite, arsenopyrite, pyrrhotite, liquid, and 
vapor can coexist. Hence, this is an in- 
variant situation where temperature, vapor 
pressure, and the compositions of the 
phases are fixed. At an invariant point in 
a three-component system five pressure- 
temperature curves of univariant equilibria 
intersect. The sequence, or order of suc- 
cession, of these P-T curves around the in- 
variant point, in either their stable portions 

ditions. The triangles depict the phase as- 
semblages stable within the divariant re- 
gions in the relevant part of the Fe-As-S 

Each of the four univariant assemblages 
containing vapor as a phase was studied 
in evacuated, sealed, silica glass tubes. In 
these assemblages the vapor pressures are 
sensitive to small temperature variations 
and the curves have low, positive slopes. 
The fifth univariant assemblage does not 






Z„\_al-_„ a 

Av ^ 



,t— — ^1 \ 


^^ i 


/ \\J \ 



-^ k 

N, \ 

S^ po + J-_^_— — — — 

A \ 




1 ' N /** 

Asp + V 

P KJ/T\ 




S V L As 

491 ±12° 
Temperature °C 

Fig. 40. Schematic, univariant pressure-temperature curves defining the phase relations and con- 
ditions surrounding the pyrite-arsenopyrite-pyrrhotite-liquid-vapor invariant point. The triangles 
depict the phase assemblages stable within the divariant regions, in the relevant part of the Fe-As-S 
system. Po, pyrrhotite; Py, pyrite; Asp, arsenopyrite; L, liquid; V, vapor. 

or metastable extensions, is determined by 
the compositions of the phases at that point, 
as has been proved and adequately illus- 
trated by Morey (1957). Reference to fig- 
ure 40, which schematically represents the 
P-T relations around the invariant point, 
will facilitate the ensuing discussion. Each 
of the univariant curves is labeled with the 
reaction for which it defines the P-T con- 

5 Since the writing of last year's report this 
reaction has been studied in detail under more 
closely controlled conditions, and this figure 
supersedes the 480° C temperature value reported 

contain a vapor phase. Hence its P-T 
curve is very steep, with the sign of the 
slope dependent on the volume change in 
the condensed reaction. This is the "upper 
stability curve," since it defines the maxi- 
mum temperature at which arsenopyrite 
and pyrite can coexist at any given con- 
fining pressure above the pressure of the 
invariant point. Along it the sum of the 
partial pressures of the elements equals the 
confining pressure. 

The upper stability curve, shown in fig- 
ure 41, has been determined at confining 
pressures up to 2070 bars. The determined 


points are: 528° ±10° C at 2070 ±10 bars, 
509° ±10° Cat 1035 ±10 bars, 494° ± 12° C 
at 210 ±35 bars, and 491° ±12° C, which is 
the invariant point at less than 1 bar vapor 
pressure, as discussed above. It is note- 
worthy that the reactions are very sluggish. 

crease measurably with increases in the 
length of the heating period at any given 
pressure. Thus the zone boundaries appear 
to be more representative of threshold en- 
ergies necessary to reverse the reactions 
than of reaction rates. At no point do the 



2? 1000 


Py + Asp — * Po + L 
Po + L —* Asp+Py 
Py + Asp do not reoct 
Po + L do not reoct 

Arsenopyrite + Pyrite 


Pyrrhotite + Liquid 




Fig. 41. The upper stability curve of the pyrite-arsenopyrite assemblage as determined in the 
presence of excess gold. The curve is drawn through the center of a zone, shown bounded by 
dashed lines, within which the reactions could not be reversed. 

Equilibrium was not attained in the runs; 
however, heating periods of 7 to 20 days 
were generally sufficient to give 10 to 30 per 
cent reaction. The trends of the reactions 
were judged on this basis. It was not pos- 
sible to reverse the reactions within a zone 
shown bounded by dashed lines in fig- 
ure 41. The width of this zone did not de- 

limits of error of the upper stability curve, 
drawn through the center of this zone, ex- 
ceed 12° C. 

The runs were performed in welded, 
collapsible, gold tubes. The solid solu- 
bility of the elemental components in the 
gold (iron being the most soluble) was 
believed to be insufficient to displace the 



upper stability curve measurably. There 
was also a considerable relative mobility 
between gold and some of the elements in 
the charges, especially arsenic, so that in 
runs of more than 4 or 5 days some of the 
material diffused through the 0.2-mm tube 
walls; in many of the runs there were 
weight losses of as much as 5 per cent of 
the initial charge. However, as in the pre- 
vious instance, the results were considered 
to be reliable because the bulk composi- 
tions remained within the region of in- 
terest. Gold tubes are not ideal contain- 
ers for studying metal-sulfur-arsenic sys- 
tems under confining pressure, but they 
are the only containers found to date that 
work at all. Projection of the curve 
through the points determined using gold 
tubes passes through the invariant point as 
determined in silica tubes. 

For the ore petrologist, the upper sta- 
bility curve defines the upper limits of 
stability of the assemblages pyrite-arseno- 
pyrite and pyrite-arsenopyrite-pyrrhotite. 
For example, at a confining pressure of 
1035 bars pyrite and arsenopyrite can co- 
exist at temperatures below 509° ±10° C 
provided that the combined partial pres- 
sures of sulfur and arsenic equal the con- 
fining pressure. It is unlikely that sulfur 
and arsenic pressures are anywhere near 
that great in the natural environment of 
ore deposition. Also the slope of the upper 
stability curve is only about 1.8° C/100 
bars. For these reasons the maximum tem- 
perature of pyrite-arsenopyrite coexistence 
in nature must be within the limit of error 
of the invariant temperature, 491° ±12° C. 
Regardless of the fact that pyrite is stable 
up to 743° C and arsenopyrite up to 702° 
C, the two minerals are not stable as a pair 
above approximately 491° C. 

R. A. Yund 

Ore deposits characterized by the occur- 
rence of native silver as well as nickel and 
cobalt sulfides, arsenides, and sulfoarse- 
nides are remarkably similar in their nickel- 
iferous phase assemblages. Deposits of this 

type have a wide geographic distribution, 
and a study of the phase relations in the 
Ni-As-S system is thus important. Such 
an investigation should make it possible to 
interpret the mineral associations, textures, 
and fine structures found in these assem- 
blages. Minerals whose compositions fall 
within the Ni-As-S system are maucherite 
(NinAss), niccolite (Nii±.,As), the NiAs 2 
polymorphs rammelsbergite and pararam- 
melsbergite, vaesite (NiS 2 ), polydymite 
(Ni 3 S 4 ), millerite (Nii-^S), heazlewoodite 
(Ni 3 S 2 ), orpiment (As 2 S 3 ), and realgar 
(AsS). Gersdormte (NiAsS) is the only 
ternary phase. Phases of Ni 5 As 2 and Ni-S6 
compositions also occur in this system, but 
these compounds have not been reported 
from nature. In addition, there is a high- 
temperature nickel sulfide (Ni 3 ±aiS 2 ) that 
cannot be quenched to room temperature. 

The phase relations were studied in 
evacuated, sealed, silica glass tubes in 
which a vapor was always present. The 
vapor pressure never exceeded about 5 bars, 
and the pressure was much lower in the 
part of the system in which a liquid or 
metallic arsenic was not a stable phase. 
The vapor volume was sufficiently re- 
stricted in each run so that the total com- 
position of the condensed phases did not 
differ measurably from the bulk composi- 

Preliminary to the investigation of the 
ternary system, the phase relations in the 
Ni-As and Ni-S systems were studied. 
The Ni-S system is discussed elsewhere in 
this report, and the phase relations in the 
Ni-As system are shown in figure 42. 
Studies of the latter system were made by 
Friedrich and Benningson (1907), and 
Heyding and Calvert (1957). Heyding 
and Calvert found that maucherite had a 
composition corresponding to the formula 
NinAss as predicted by Peacock (1940). 
Some evidence indicating incongruent 
melting of this phase was also presented. 
The present study has confirmed the incon- 
gruent melting of maucherite (to niccolite 
plus a liquid) at 829° C. In addition, the 



experimental work revealed a small change 
in the cell dimensions with arsenic con- 
tent of maucherite. Experiments at 700° C 
showed that the value of </ 2 oo of maucher- 
ite in equilibrium with Ni 5 As 2 is 0.0012 A 
smaller than that of maucherite in equilib- 
rium with NiAs. This small difference is 
real and indicates a very narrow range in 
the arsenic content of maucherite. Subse- 

to determine the nature of this solid solu- 
tion. On the nickel side of stoichiometric 
niccolite (56.07 weight per cent arsenic) 
the solid solution is due to excess nickel 
atoms in the structure, and the formula is 
represented by Nii +a As. On the arsenic 
side of stoichiometric niccolite the solid 
solution is due to nickel omission, which 
is represented by the formula Nii-aAs. 


Weiqht per cent As 
Fig. 42. T-X diagram of the Ni-As system. Vapor is present in all assemblages, and the pressure 

of the system is not constant. 

quent work has shown that the composi- 
tional variation is less than 0.5 weight per 
cent even at 800° C. Maucherite is tetrag- 
onal and when formed in equilibrium with 
Ni 5 As 2 has cell dimensions of a = 6.870 ± 
0.001 A and c=21.81 ±0.01 A. 

Niccolite has a range in composition at 
800° C from 55.4 to 57.20 weight per cent 
arsenic. Measured densities and densities 
calculated from unit cell dimensions of 
various niccolite compositions were used 

NiAs is hexagonal, and unit cell dimen- 
sions of stoichiometric niccolite and nicco- 
lite with the maximum nickel deficiency 
at 700° C are a=3.618± 0.001, c= 5.034 ± 
0.001 A; and a = 3.616 ±0.001, c=5.025± 
0.001 A, respectively. As shown in figure 
42, the slope of the niccolite solvus is nearly 
vertical below 700° C. Therefore, the com- 
position of niccolite formed in equilibrium 
with rammelsbergite (aNiAs 2 ) or para- 
rammelsbergite (3NiAs 2 ) is not suffi- 



ciently variable to be used as a geologic 

The determination of the liquidus be- 
tween NiAs and NiAs 2 is complicated by 
the dissociation of rammelsbergite to nic- 
colite plus a vapor. The eutectic between 
niccolite and rammelsbergite is at 853° 
±2° C and a composition of 64.0 ±1.0 
weight per cent arsenic. The melting tem- 

blages (three condensed phases plus vapor) 
are shown as clear areas; three-phase bi- 
variant assemblages by narrowly spaced, 
lightly ruled lines; and two-phase trivari- 
ant regions are solid black. As realgar and 
orpiment melt slightly above 300° C they 
do not appear on the ternary diagrams in 
figures 43 and 44. The only condensed 
phases stable above 500° C between arsenic 

/3NiAs 2 

Wt. per cent 

Fig. 43. Phase relations in the Ni-As-S system at 500° C. Vapor pressure varies with the assem- 
blage, hz, heazlewoodite; va, vaesite; ma, maucherite; nc, niccolite; prm, pararammelsbergite; gf, 
gersdorffite; L, liquid; V, vapor. Stable assemblages as shown are: (1) (Ni,As)-hz-V, (2) (2V7,As)- 
Ni 5 As„-hz-F, (3) Ni 5 As 2 -hz-F, (4) Ni 5 As 2 -ma-F, (5) Ni 5 As 2 -ma-hz-F, (6) ma-hz-aNi 7 S 6 -F, 
(7) ma-aNi 7 S 6 -c<NiS-F, (8) ma-aNi(S,As)-V, (9) ma-nc-aNi(S,As)-F, (10) ma-nc-F, (11) nc- 
prm-gf-F, (12) nc-gf-F, (13) nc-ctNi(S,As)-gf-F, (14) aN^ ffi (5,As)-gf-F, (15) aNi^S-gf- 
va-F, (16) va-L-F, (17) gf-va-L-F, (18) gf-L-F, (19) As-L-gf-F, (20) prm-As-gf-F. 

perature of NiAs2 was not determined, 
but it must be above 1040° C since no 
melting was observed at this temperature. 
Preliminary data indicate that the poly- 
morphic inversion of pure NiAs 2 is at 
approximately 550° C. 

The phase relations in the ternary sys- 
tem at 500° C are shown in figure 43; iso- 
thermal sections at higher temperatures, 
in figure 44. Four-phase univariant assem- 

and sulfur are a liquid and metallic ar- 
senic. The amount of nickel soluble in this 
liquid has not been determined. L. A. 
Clark (1959) found that approximately 
0.1 per cent iron is soluble in an arsenic- 
sulfur liquid at 600° C, and it is reasonable 
to expect the solubility of nickel to be about 
the same as that of iron. 

Niccolite and Nii-*S form a complete 
solid solution series above 600° C, whereas 



a solvus with its crest at approximately 
50 weight per cent NiS appears slightly 
below 600° C. The width of the two-phase 
region greatly increases with decreasing 
temperature (see figs. 43 and 44#). At 
500° C the two phases, Ni(^^S) and 
Ni(S,As), have compositions of 3.9 and 
32.2 weight per cent sulfur, respectively. 

The limits of sulfur solubility in gersdorff- 
ite at various temperatures have not been 
accurately determined, owing to slow reac- 
tion rates. (Equilibrium was not attained 
at 700° C after 3 months and two regrind- 
ings of the material.) A time versus com- 
position study of this part of the gersdorff- 
ite solid solution series is in progress. Equi- 

Fig. 44. Isothermal sections of the Ni-As-S system. Vapor is present in all assemblages, and the 
pressure varies with the assemblage. 

The increase in the extent of solid solution 
perpendicular to the line from NiAs to NiS 
at temperatures above 600° C is shown in 
figure 44£ to d. This increase is due to the 
formation of a more nickel-deficient 
Nii-.z(As,S) solid solution. 

Gersdorffite has a wide range in compo- 
sition with respect to the As/S ratio but 
shows no variation in the Ni/As4-S ratio. 

librium was easily attained, however, in 
the synthesis of gersdorffites along the 
NiAsS-rammelsbergite join (As/S ratios 
greater than 1). The maximum arsenic 
content of gersdorffite at 700°, 600°, and 
500° C is 64.1, 56.8, and approximately 52 
weight per cent, respectively. Gersdorffite 
has a cubic lattice, and the length of the 
cell edge as a function of composition is 


shown in figure 45. It is interesting to 
note that the change in a is linear through- 
out the compositional range of As/S ratios 
greater than 1 and that the value of a in- 
creases from NiAsS toward NiAs2. These 
facts suggest that the solid solution is pri- 
marily due to the substitution of the larger 
arsenic ion for the smaller sulfur ion. 

Rammelsbergite and maucherite have a 
limited sulfur solubility at high tempera- 
ture. The maximum sulfur content of ram- 
melsbergite at 700° C is 1.2 weight per cent. 

high-temperature nickel sulfide Ni3±sS2 
could not be retained on quenching; it was 
possible, however, to quench this phase 
when it was formed in equilibrium with 
maucherite. Presumably the substitution of 
some arsenic for sulfur stabilizes the struc- 
ture at lower temperatures. Although this 
phase may vary considerably in composi- 
tion, it is represented by a point on the dia- 
grams in figure 44. The formula Ni 3 ±a:S2 
has been used to denote it and to dis- 
tinguish it from heazlewoodite — the low- 

Wl per cent As 
6120 55 87 

"1 1 1 r 

NiAs 2 Nlflsb 

Wl per cent NiAsS 

Fig. 45. a as a function of gersdorffite composition. Calculation of a based on the (311) reflection. 

Rammelsbergite is orthorhombic, and pure 
NiAs 2 has unit cell dimensions of a = 4.756, 
£=5.793, and c — 3.544 A, whereas ram- 
melsbergite containing 1 weight per cent 
sulfur has unit cell dimensions of a = 4.739, 
£=5.781, and c- 3.558 A. At 700° C mau- 
cherite dissolves approximately 3.5 weight 
per cent sulfur, and has unit cell dimen- 
sions (tetragonal) of a = 6.842 and c — 21.73 
A. The maximum sulfur content of ram- 
melsbergite, or pararammelsbergite, and 
maucherite has not been determined at 
lower temperatures. 
It was previously mentioned that the 

temperature, stoichiometric Ni 3 S 2 phase. 
(See discussion of the Ni-S system else- 
where in this report.) 

Ternary phase assemblages at various 
temperatures are shown in figures 43 and 
44. At 500° C the assemblages heazlewood- 
ite-maucherite-aNiS-vapor and heazle- 
woodite-maucherite-aNirSs-vapor are sta- 
ble. Ni3±a;S2 becomes the stable phase 
between 525° and 550° C, depending on 
its composition. At 570° ±2° C aNi T S c 
breaks down peritectoidally to Ni3±^S 2 and 
aNiS, giving rise to the new assemblage 
Ni3±aS2-maucherite-aNiS-vapor. This as- 


semblage is stable to about 600° C, at 
which temperature a composition along 
the NiS-niccolite solid solution series be- 
comes the stable phase in equilibrium with 
maucherite and Ni 3 ±tfS 2 . The tie lines 
from these phases meet at a common point 
on the NiS-niccolite solid solution series. 
The composition of this point moves to- 
ward NiAs with increasing temperature 
(fig. 44c and d). In addition, the two uni- 
variant assemblages, maucherite-Ni(/4i', 
S)-Ni( l S,As) -vapor and gersdorffite-Ni- 
(^i , ,S)-Ni(5',As)-vapor, also disappear at 
600° C, at which temperature niccolite and 
aNii-zS forms a complete solid solution 
series. The composition of gersdorffite in 
the last-mentioned assemblage is NiAsS. 

Stable assemblages shown in figure 43 
closely correspond to most natural assem- 
blages, except that Nii-^S always occurs in 
nature as the low-temperature polymorph 
millerite. This is readily explained because 
even at room temperature aNiS inverts to 
millerite in a few months. In addition, the 
maximum arsenic content of millerite is 
much less than that of aNiS. The variation 
with temperature of the sulfur content of 
niccolite occurring in either of the assem- 
blages niccolite-millerite-gersdorffite or 
niccolite-millerite-maucherite is a poten- 
tial geologic thermometer. Unfortunately 
these assemblages are not common in na- 
ture, and the application of the niccolite 
solvus for temperature determinations is 
greatly limited. An assemblage more com- 
mon in nickel deposits is rammelsbergite 
or pararammelsbergite-niccolite-gersdorff- 
ite, and, inasmuch as the composition of 
gersdorffite varies markedly with tempera- 
ture, this solvus offers a possible method 
of determining geologic temperatures. The 
effect of substitution of other elements in 
the gersdorffite structure, particularly Co, 
Fe, and Sb, as well as the effect of pressure 
on the As/S ratio must be known, how- 
ever, before quantitative use can be made 
of this solvus. If the results of a current 
study at lower temperatures indicate that 
the solvus has a moderate slope, the effect 
of the aforementioned variables on the 

As/S ratio can be investigated. At present 
the application of this solvus appears prom- 
ising, for natural gersdorffites occurring 
with rammelsbergite or pararammelsberg- 
ite and niccolite do have a variation in their 
As/S ratios. 


E. H. Roseboom, Jr. 

In the two preceding reports (Year 
Books 56 and 57) a reconnaissance of the 
arsenic-rich part of the system Co-Ni- 
Fe-As at temperatures of 600° to 800° C 
was described. A literature review of the 
ore deposits pertaining to this system indi- 
cates that the minerals and mineral assem- 
blages commonly are the same as those 
synthesized in the laboratory. In some 
cases, however, there were disagreements 
which indicate that some of the phase rela- 
tions found in ores cannot be inferred from 
extrapolations of the synthetic relations 
obtained at 600° to 800° C. 

Of one hundred references critically ex- 
amined, approximately sixty provided per- 
tinent information. The references had to 
be re-evaluated because of changes in no- 
menclature and because some minerals 
have been discredited. Proof of equilibrium 
in nature cannot be obtained. The only 
criterion of equilibrium used was actual 
physical contact between the minerals. 

The arsenides of cobalt, nickel, and iron 
occur together most frequently in hydro- 
thermal deposits that have come to be rec- 
ognized as a world-wide type. The largest 
of these are the Erzgebirge in Saxony and 
Bohemia, Cobalt in Ontario, and Great 
Bear Lake in the Northwest Territories, 
but at least forty smaller deposits have 
been described. Silver, bismuth, and oc- 
casionally arsenic are present as native 
metals. The arsenides sometimes encrust 
bismuth crystals, indicating deposition be- 
low the melting point of bismuth at 271° 
C. Sulfides are not abundant, and sulfates 
are rare. 

In addition to the Co-Ni-Ag type of 
deposit, skutterudites have been found in 



high-temperature and contact metamorphic 
deposits, loellingite in pegmatites and con- 
tact metamorphic deposits, and niccolite 
with chromite in peridotites and norites. 

Figure 46 represents the probable phase 
relations between minerals in ore deposits 
for the arsenic-rich part of the system 
Co-Ni-Fe-As. This figure was constructed 
by taking the phase relations as known 
at 600° to 800° C and modifying them to 

The most nickel-rich loellingite analyses 
are under 40 per cent NiAs2, and most 
analyses contain less than 3 per cent N1AS2. 
Eight references to coexisting rammels- 
bergite and loellingite were found, sug- 
gesting that occurrences of this pair may 
be common enough so that the nickel con- 
tent of loellingites stable with rammels- 
bergite can be used as a thermometer. 

Three references were found for the loel- 

Section through 
the system on 
the RAs 2 plane 

As = Arsenic 

: loellingite 
mh = moucherite 
nc = niccolite 
rm*rammelsbergite or 

parara mmelsbergite 
sf = sofflorite 

( ) indicate that enclosed phases 
do not lie on plane of projection 

(nc + sk) 
+ sf 



Fig. 46. Co-Ni-Fe-As phase relations for natural assemblages of minerals. 

fit the assemblages observed in natural 

Ni-Fe-As. The part of this system more 
arsenic rich than NiAs 2 -FeAs 2 has been 
studied experimentally. The same mineral 
assemblages occur naturally, and that part 
of figure 46 is very similar to the phase 
diagram of the synthetic system. The 
solvus between loellingite and rammels- 
bergite should become wider at lower 
temperatures, however. At 750° C the 
nickel-richest phase on the loellingite side 
of the solvus is about 67 per cent NiAs 2 . 

lingite-niccolite assemblage but none for 
loellingite-maucherite. This join is postu- 
lated because FeAs and phases higher in 
Fe are not known as minerals. Joins al- 
ternative to loellingite-maucherite would 
involve niccolite and FeAs or phases higher 
in Fe content. 

Fe-Co-As. The arsenic-rich part of this 
system at 600° to 800° C is similar to 
figure 46, except that, as ten references to 
coexisting loellingite and safflorite were 
found, it seems likely that this series is 
not complete in natural deposits. Analyses 


of loellingites extend from FeAs 2 to about 
25 per cent CoAs 2 . Three analyses of mix- 
tures of safflorite and loellingite indicate 
that total compositions of the pairs are 
around the 1:2 ratio of Co:Fe. Therefore, 
this composition should lie in the break in 
the series. Nine references support the 
loellingite-skutterudite assemblage, and 
four report the loellingite-safflorite-skut- 
terudite assemblage. 

Co-Ni-As. The part of this system found 
in natural deposits differs from the system 
studied experimentally in that the series 
between CoAs 2 and NiAs 2 is incomplete 
and a join exists between niccolite and 
skutterudite (ten references). The pair 
niccolite-skutterudite appeared at 800° C 
in runs with compositions between CoAs 2 
and NiAs 2 but disappeared with grinding 
and reheating, and diarsenide solid solu- 
tions formed instead. Heating the diarse- 
nide solid solutions at 600° C for 3% 
months failed to break them down. 

It seems likely that at lower temperatures 
the solvus between rammelsbergite and 
loellingite in the system Ni-Fe-As extends 
farther into the Co-Ni-Fe diarsenides un- 
til it reaches the series CoAs 2 -NiAs 2 and 
interrupts it. At a still lower temperature, 
the cobaltian safflorite-rammelsbergite pair 
in the series CoAs 2 -NiAs 2 becomes unsta- 
ble and niccolite-skutterudite becomes 

Co-Ni-Fe-As. The situation inside the 
Co-Ni-Fe-As tetrahedron is difficult to 
depict. There are four four-phase regions 
which appear as distorted tetrahedra: (1) 
udite, (2) loellingite-rammelsbergite-saf- 
florite-skutterudite, (3) loellingite-ram- 
melsbergite-safflorite-niccolite, (4) loel- 

The first four-phase region is intersected 
by the section through the tetrahedron on 
the diarsenide plane (fig. 46). The next 
two regions have three corners of their 
tetrahedra (loellingite, rammelsbergite, saf- 
florite) on the diarsenide plane and the re- 
maining corner above or below this plane. 
The fourth region has three corners (loel- 

lingite-maucherite-niccolite) near the Fe- 
Ni-As side of the tetrahedron and the 
fourth corner inside at some composition 
of safflorite. 

The most likely parts of the system to 
yield useful geothermometers are the ram- 
melsbergite-loellingite solvus and the break 
in the loellingite-safflorite series. 


Application of the Pyrrhotite X-Ray De- 
terminative Curve to Natural Pyrrhotites 
R. G. Arnold and Laura Reichen 6 

Last year's report described a prelimi- 
nary attempt to use the experimentally 
determined pyrrhotite X-ray spacing versus 
composition curve for estimating the com- 
position of natural pyrrhotites. It was con- 
cluded on the basis of analyses of six sam- 
ples for iron that the curve could be used 
if the pyrrhotites contained less than 0.4 
weight per cent combined nickel and 

The applicability of this curve to natural 
pyrrhotites was demonstrated by means of 
fifteen additional natural pyrrhotites, which 
were chemically analyzed for both iron and 

Iron content determined by chemical 
analyses and by the X-ray method on the 
same samples differed by only 0.21 atomic 
per cent iron in one sample and by less 
than 0.15 atomic per cent iron in the re- 
maining ones. These differences are less 
than the combined uncertainty of the two 
methods (0.13 atomic per cent iron in the 
X-ray measurement and 0.15 atomic per 
cent iron or more in the chemical analy- 
ses) ; consequently, the iron contents deter- 
mined by both methods in each case are 
essentially the same. The compositions of 
the pyrrhotites range from 47.98 to 46.68 
atomic per cent iron. 

Nickel, cobalt, manganese, and copper, 
which are the commonest elements found 
in solid solution in the pyrrhotites, were 

G U. S. Geological Survey, publication approved 
by the Director. 


determined by the X-ray spectrograph. The 
maximum individual concentrations of 
nickel, cobalt, and copper found were 0.47, 
0.16, and 0.15 weight per cent, respectively. 
Manganese was always less than 0.01 
weight per cent. The maximum combined 
concentration of nickel, cobalt, and copper 
was 0.58 weight per cent. Experiments 
showed that 2 weight per cent nickel, 2 
weight per cent cobalt, enough copper to 
saturate pyrrhotite at 600° C (<0.85 
weight per cent) separately, and 1 weight 
per cent nickel, 1 weight per cent cobalt, 
and 0.6 weight per cent copper combined 
may be present in pyrrhotite without caus- 
ing a measurable shift in the ^(102) values. 
The concentration of impurities in our 
samples is thus insufficient to affect the 
</(102) values. 

Temperatures of Formation of Coexisting 

Pyrrhotite, Sphalerite, and Pyrite from 

Highland Surprise Mine, Idaho 

R. G. Arnold, R. G. Coleman, and 
G. C. Fryfylund 6 

The Highland Surprise Mine is located 
in Shoshone County, Idaho, and lies within 
the Coeur d'Alene lead-zinc-silver district. 
Two of the three producing veins, nos. 1 
and 2, are of particular interest; they strike 
about N 65° W, dip nearly vertically, have 
an average stope length of 300 and width 
of 6 feet, a maximum pitch length of 700 
feet, and a rake of 70°. The veins were 
formed by the replacement of the Pritchard 
formation, a member of the Precambrian 
Belt series. This formation is a banded 
sedimentary sequence consisting princi- 
pally of quartz, feldspar, and sericite, with 
rock fragments and heavy accessory min- 

The veins consist of one or more of the 
sulfides pyrrhotite, pyrite, arsenopyrite, 
sphalerite, galena, and chalcopyrite, to- 
gether with quartz, sericite, a little chlorite, 
and very minor amounts of ankerite and 
siderite. Samples of vein material both of 
ore grade and non-ore grade were collected 
over a vertical and lateral interval of 1650 
and 2000 feet, respectively. 

Good ore consists of a mixture of galena 
and sphalerite, either of which may pre- 
dominate, together with pyrrhotite and 
one or more of the other sulfides, and fre- 
quently up to 30 or 40 per cent of nonsul- 
fide material. Non-ore-grade material con- 
sists principally of pyrrhotite with pyrite 
and arsenopyrite and lesser amounts of one 
or more of the other sulfides. 

Pyrite occurs with pyrrhotite in three 
ways: in massive concentrations that ap- 
pear to have been fractured; as scattered 
euhedral and rounded grains up to 3 mm 
across; and as minute veinlets or blebs up 
to 0.1 mm long and 0.03 mm wide. Pyrite 
appears to have been deposited in two gen- 
erations, one younger, and the other older, 
than pyrrhotite. 

The depositional relations between sphal- 
erite and pyrrhotite are not clear-cut. On 
the basis of textural relations and the rela- 
tive position of the pyrrhotite and sphaler- 
ite masses in the veins it appears that they 
were deposited at about the same time. No 
exsolution textures of either chalcopyrite 
or pyrrhotite in sphalerite were observed. 

Estimates of the crystallization tempera- 
ture of pyrrhotite were obtained from each 
of 62 samples in the following way. The 
iron content of each pyrrhotite was deter- 
mined by the X-ray method (described 
above), and the corresponding temperature 
was obtained from the pyrite-pyrrhotite 
solvus curve presented in last year's report 
(fig. 27, p. 219). 

The estimated temperatures of forma- 
tion ranged from 370° to 492° C. These 
values are summarized graphically in fig- 
ure 47 in 5° C intervals. The error in each 
temperature estimate is about 25° C, which 
corresponds to the error in the iron deter- 
minations. About 54 per cent of the tem- 
perature estimates are located in the tem- 
perature interval 450° to 464° C, about 20 
per cent are less than 450° C, and 26 per 
cent are above 464° C. A plot of the 62 
pyrrhotite temperatures at their respective 
sample locations on a vertical longitudinal 
section showed a mixture of high and low 
temperatures suggestive of more than one 


generation of mineralization. Comparison 
of the temperatures of the ore-grade sam- 
ples with those of the non-ore-grade sam- 
ples showed no significant differences. 

No correction for the confining pressure 
prevailing during crystallization need be 
applied to these pyrrhotite temperature 
estimates for pressures up to 2000 bars, con- 
trary to the results reported in Year Book 
56. 7 Spectrographic analysis of 28 pyrrho- 
tites from widespread locations in the mine 
shows that the maximum concentration of 
combined nickel, cobalt, and copper in pure 
samples is 0.145 weight per cent. The con- 
centration of impurities in these pyrrho- 
tites is too low to have a measurable effect 








I - 


i , n , : 

; r^n « 


380 400 420 440 460 480 500 

Temperature °C — *• 

Fig. 47. Estimated pyrrhotite-pyrite tempera- 
tures of formation summarized graphically in 
5° C intervals. 

on either the d(102) values or the position 
of the curve representing the composition 
of sulfur-saturated pyrrhotite. 

The crystallization temperature of sphal- 
erite coexisting with pyrrhotite from 14 
samples was determined by the method de- 
veloped by Kullerud (1953). These tem- 
peratures range from 375° to 490° C, 
closely matching the range shown by the 

7 It was reported that the curve representing 
the composition of pyrrhotite that can coexist 
with pyrite shifted significantly toward iron- 
richer compositions with confining pressures of 
up to 2000 bars. Using a bracketing technique it 
was found that pressures of up to 2000 bars 
below 670° C did not measurably shift the curve. 
The former results were found to be in error due 
to the exsolution of pyrite from pyrrhotite during 

pyrrhotite temperatures. According to Kul- 
lerud's work a positive temperature correc- 
tion of about 25° C/1000 bars total pres- 
sure should be applied to these estimates. 

The iron contents of the sphalerites were 
determined by standard wet chemical 
means to an accuracy of about 0.15 weight 
per cent. The corresponding error in the 
temperature estimates is about 8° C. Cad- 
mium and manganese in solid solution are 
less than 0.44 and 0.068 weight per cent, 
respectively; the largest concentration of 
cadmium and manganese combined is 
0.465 weight per cent. Kullerud's studies 
show that no correction to the temperature 
values is necessary for these small concen- 
trations of impurities. 

Comparison of the sphalerite and pyr- 
rhotite temperatures obtained from the 
same samples indicates that 13 of the sphal- 
erite temperatures are from 2° to 92° C 
lower than that of the coexisting pyrrho- 
tite, whereas the fourteenth sphalerite tem- 
perature is 9° C higher than the tempera- 
ture estimate of the coexisting pyrrhotite. 

The temperatures obtained by the two 
methods agree fairly well in view of the 
probable errors in the two methods (±25° 
C for pyrrhotite-pyrite and ±10° C for 
pyrrhotite-sphalerite), the apparent tem- 
perature discrepancies caused by rock pres- 
sure (no correction on pyrrhotite-pyrite 
solvus, but about 25° C/1000 bars on sphal- 
erite-pyrrhotite solvus), and the additional 
uncertainties posed by comparison of sam- 
ples from possibly differing generations of 
ore formation. 


S. P. Clark, Jr. 

The development of the sphalerite geo- 
thermometer permits a fresh attack on 
thermal problems associated with certain 
types of ore deposition. As is usual in geo- 
logic applications of the theory of heat 
conduction, the problem must be some- 
what idealized for mathematical conven- 
ience, and highly accurate results cannot 
be expected. Nevertheless some useful lim- 


its can be obtained from analysis based on 
the temperatures recorded in ore deposits. 

In the Wellington Mine in Colorado, 
Kullerud found that the temperature of 
deposition, determined from the iron con- 
tent of sphalerite coexisting with pyrrho- 
tite, varied linearly from 410° to 500° C 
in a vertical distance of 300 feet. Hence 
the thermal gradient along the vein is 
about 1000°C/km. This is much larger 
than the usual geothermal gradient in crys- 
talline rocks (ca. 20°C/km). The follow- 
ing analysis indicates that, for any assumed 
duration of deposition, a rather definite 
upper limit to the rate of flow of the ore- 
forming fluid can be set. If this limiting 
rate is exceeded, the observed thermal 
gradient could not have persisted through- 
out the time of deposition of the ore. This 
relation, in conjunction with further 
thermal and chemical considerations, per- 
mits less definite limits to be assigned to 
the concentration of the ore-forming fluid. 

The normal geothermal gradient is so 
much less than the gradient in the vein 
that it can be neglected. We choose our 
temperature scale so that the initial tem- 
perature of the country rock is zero. The 
ore-bearing vein is treated as a slab, and 
coordinates are chosen so that one of its 
faces is the plane 2 = 0. It is assumed to ex- 
tend indefinitely in the y direction. After 
time £ = fluid flows along the slab with 
uniform velocity in the direction of posi- 
tive x. Within the slab the temperature 
depends only on x and t, and the plane 
x = is maintained at a temperature To. 

A final approximation required to obtain 
an analytic solution is to assume that the 
thermal conductivity of the wall rock is 
equal to K in a direction perpendicular to 
the vein and to zero parallel to it. That is, 
heat flow in the x direction is neglected. 

For convenience the notation used in the 
analysis is collected here. T'(x, t) = tem- 
perature in the vein, T(x, z, t)— tempera- 
ture in the wall rock, To = temperature of 
the fluid at x = 0, U = velocity of flow of 
the fluid, c = heat capacity of the fluid, 
p = density of the fluid, k = thermal diffu- 

= --^f- x>0,z>0,t>0 

sivity of the wall rock, K = thermal con- 
ductivity of the wall rock in the z direction, 
21 = width of the vein, \ = K/pcl. 
We must solve the equation 

9 2 T_ 1 dT 

dz 2 k dt 

subject to the conditions 

T=T',x>0,t>0,z = 

dT dT dT' 
k=r = ^ + U ^- x>0,t>0,z = 
oz ot ox 


T = 0,x>0,z>0,t=0 

T' = T o ,x = 0,t>0 

The solution to this problem, which is 
given by Carslaw and Jaeger (1947), can 
be expressed in terms of the complemen- 
tary error function, denoted erfc y. It is 
defined by the relation 

2 C 00 
erfc y — — = / e " m2 du 

Vtt' y 

and tables of this function are readily 
available. The solution for the tempera- 
ture in the vein is 


— -=errc i -7— 
To [_Bk[*- 

and the temperature in the wall rock is 
given by the expression 


r kx/U I 
[_{4k[;- (x/U) ]}*J 

r f r {kx/u)+z I 

^o L{4k[/- (x/C/)]>* J 


A typical plot of the temperature in the 
vein is given in figure 48. 

The width of the vein in the Wellington 
Mine is somewhat less than 1 meter. Tak- 
ing K= 0.005 cal/cm sec °C (a representa- 
tive value for quartz-monzonitic rocks), 
p = l g/cm 3 and c=\ cal/g °C (values 
appropriate to liquid water), and 1=50 
cm, we find ^ = 10 -4 cm/sec. If the den- 
sity and heat capacity of molten pyrrhotite 
are chosen instead, ^ is roughly 4/3 this 

The quantity To is the difference be- 
tween the undisturbed temperature of the 
rock and the "initial" temperature of the 


fluid. The melting point of the material 
in the vein is likely to be about 650° C, 
and the country rock must have originally 
been hotter than 0° C. Hence To cannot 
exceed about 600° C, especially since the 
ore solution may have moved a consider- 
able distance from its source to the part 
of the vein in which the observations were 

Temperatures in the vein were calcu- 
lated for values of \ ranging from 5 X 10~ 5 
to 10" 3 and for times ranging from 10 6 to 
10 12 sec (12 days to 30,000 years) with 
k=0.01 cm 2 /sec. Within this range of the 
parameters a nearly constant gradient is 
obtained if U is taken proportional to t~ h . 
The above values of To and the gradient 

V 0, 


X, meters 

Fig. 48. Temperature along the vein for 
^ = 10~ 4 and two values of time and rate of flow. 

in the vein lead to the approximate re- 

In the 100-meter interval from which 
samples were taken there are 10 4 cm 3 of 
ore for each square centimeter of vein 
measured in the direction of fluid move- 
ment. This corresponds to 5 X 10 4 g if the 
density of the ore is 5 g/cm 3 . The amount 
of fluid which crosses each square centi- 
meter of vein is pUt, which by the above 
relation for U is roughly equal to 50ptK If 
p = 1 g/cm 3 , this relation yields 5 X 10 T g in 
10 12 sec or 5X10 4 g in 10 G sec. The latter 
result, which corresponds to a concentra- 
tion of 100 per cent, is not realistic, since 
such a solution would have a density con- 
siderably greater than unity. All we can 
conclude is that if deposition is to take 

k = IO-" 






= IO l2 ,U = IO" 4 




t = l0 6 ,U = IO" 

1 A \ 







i i i nil i 

1 Mill 

Vj v 


place in 10° sec a highly concentrated solu- 
tion is required. 

Concentrations of sulfides of the order of 
10 per cent in the solution are probably 
excluded by an argument based on the 
heat of precipitation. In view of the low 
solubility of sulfides in any known solution 
at low temperatures, it is probable that a 
strongly positive temperature coefficient of 
solubility is required if such concentrations 
are to become possible. But this implies 
that the process of precipitation is strongly 
exothermic, which in turn means that the 
solution can heat the wall rock without 
being cooled itself. It is then difficult to 
understand how a thermal gradient of 
1000°C/km could be set up along the vein. 
In more dilute solutions heat of precipita- 
tion raises no difficulty. 

The present solution gives no upper 
limit to the duration of deposition, prob- 
ably because heat conduction in the direc- 
tion of fluid flow has been neglected. For 
times less than 10 10 sec this approximation 
is valid, but for longer times it may be 
expected to break down. This leads to an 
overestimate of the thermal gradient in 
the vein, since heat flow in the wall rock 
tends to reduce temperature differences. 
Rough estimates suggest that conduction 
would eliminate the gradient along the 
vein in 10 14 sec, and that the upper limit 
to the duration of deposition cannot greatly 
exceed 10 12 sec. This implies that the metal 
content of the solution is 0.1 per cent or 

The results of a calculation of this sort 
are no better than the material constants 
used in obtaining them, and it is necessary 
to examine the values adopted for the 
thermal parameters more closely. The 
density and heat capacity of the fluid enter 
the problem only through the parameter 
f{, and a variety of values have been tried 
without great effect on the results. The 
product pc for liquid sulfides seems to be 
close to that of liquid water, judging from 
the data for pyrrhotite. This product could 
be smaller in a vapor, but experience with 



the solubility of solids in gases suggests 
that the requisite concentration of sulfides 
could not possibly be attained in a low- 
density vapor. 

The values adopted for the thermal 
conductivity and diffusivity of the wall 
rock are typical of granitic or granodi- 
oritic rocks at moderately high tempera- 
tures. K is unlikely to be much less than 
0.005 cal/cm sec °C, but if fluid from the 
vein permeated the wall rock, the effective 
conductivity could be greater. The ratio 
of the amount of heat carried by fluid to 
that conducted from the vein is roughly 
cwD/K, where w is the number of grams 
of fluid escaping per square centimeter of 
wall per second, D is the distance the fluid 
travels, and the other symbols have their 
previous meanings. With c = l cal/g °C 
and the above value of K, we find that 
the two processes are equally important 
when u/D = 5x\0~ 3 g/cm sec. 

From Darcy's law, the pressure drop in 
the wall will be wDv/p\x, where v is the 
viscosity of the fluid in centipoises and u 
is the permeability of the wall rock in 
darcys. The last figure can hardly exceed 
10" 4 millidarcy, which with a fluid having 
the viscosity of water implies a pressure 
drop in excess of 10 kilobars. No reason- 
able adjustment of the constants used in 
obtaining this result is likely to make it 
acceptable, and we conclude that transport 
of heat by percolating solutions is unim- 
portant except possibly along localized 
fissures. This conclusion is consistent with 
the scarcity of disseminated sulfides in the 
wall rock and the lack of wall-rock altera- 
tion along these particular veins. 

The analysis of temperatures in this vein 
leads to a concentration of 0.1 to 10 per 
cent sulfides in the ore-forming fluid — a 
surprisingly sharp result, considering the 
nature of the method and the fact that the 
vein was originally sampled for other pur- 
poses. There is no obvious reason to doubt 
the present results, but they could be ac- 
cepted with greater confidence if some 
further studies were made. We have in- 
sufficient information about the range of 

temperatures at a fixed level in the mine, 
for example. Information about the tem- 
peratures in the wall rock as a function of 
distance from the vein (which we lack) 
could give a more precise estimate of the 
time required for deposition. The conclu- 
sions are tentative, but the present results 
are considered promising enough to justify 
further sampling in the Wellington Mine 
and elsewhere. 


E. H. Roseboom, Jr. 

The vapor pressures over sulfides or sul- 
fide assemblages are detectable even at 
low temperatures. In a closed sulfide sys- 
tem the vapor consists mainly of sulfur 
and forms if space is available. If it con- 
tains much metal or departs from ideality, 
the sulfur vapor can be treated in terms of 
the activity of sulfur. In the case of a 
natural sulfide in equilibrium with the 
depositing solution, the activity of sulfur 
should be the same for a particular phase 
or assemblage in the solution as it is in 
the vapor phase in the dry synthetic sys- 
tem. Thus the vapor pressure and its com- 
position over a particular sulfide or assem- 
blage of sulfides can be used to determine 
the activity of sulfur in the coexisting 
solution if the temperature at the time of 
deposition is known. The activity of sulfur 
in turn sets certain limits on the composi- 
tion of the ore solution and the ionic 
species occurring in it (see Barnes, Year 
Book 57). Also, if the vapor pressure of 
sulfur can be established from certain 
mineral compositions or assemblages, the 
usefulness of some existing sulfide geo- 
thermometers can be greatly extended, as 
shown by Barton and Kullerud (Year 
Book 57). 

Vapor pressures over a number of sul- 
fides have been measured using silica spiral 
manometers, dew-point determinations, or 
the analysis of hydrogen-hydrogen sulfide 
mixtures in equilibrium with the sulfide. 
In some cases such as pyrite-pyrrhotite, the 
published data vary widely at lower tern- 


peratures and pressures. The dew-point 
method has been criticized because the 
strong temperature gradients across the 
vapor phase make equilibrium doubtful 
among the various sulfur species, S2, S6, 
and Ss, in the vapor. Moreover, metals 
may be transported in the vapor and be- 
come dissolved in the condensed sulfur, 
altering its condensation vapor pressure 
from that of pure sulfur. The silica spiral 
manometers are satisfactory but delicate 
and difficult to make, where high sensi- 
tivity is required. 

Most of the difficulties of these methods 
can be avoided if a suitable manometer 
liquid of low vapor pressure can be found 
that will not react with sulfur. At this Lab- 
oratory preliminary measurements have 
been made on the vapor pressure of sulfur 
using a silica glass U tube containing a 
eutectic mixture of LiCl and KCl. This 
mixture melts at 350° C. Because of the 
low density of the salt melt, the displace- 
ment of the salt column is about eight 
times that of the mercury at any given 
pressure. As the salt column can be read 
reproducibly to 0.1 mm by means of a 
transit, the lower limit of pressure meas- 
urement should be approximately 0.01 mm 

In practice, the sample is placed in the 
closed end of the manometer and the mix- 
ture of solid salts is poured into the U tube. 
The salt is melted by means of a nichrome 
heating coil wound around the manom- 
eter. The molten salt is poured into a side 
tube with a dead end, and the system is 
evacuated. When the system is at a suf- 
ficiently low vacuum, the molten salt is 
poured back into the U tube and left for 
a time. If the system has been sufficiently 
degassed, the salt levels should remain 
constant. If not, the salt can be poured 
off and degassing continued. 

Then the pressure in the system outside 
the manometer is changed several times, 
and the changes in levels of the molten 
salt are calibrated against the changes in 
levels in a manometer containing octoil. 

Finally a furnace or a bath of molten salt 
is used to heat the sample to the desired 
temperatures. When the pressures are less 
than 10 mm Hg, the manometer is read 
directly with the system outside the ma- 
nometer evacuated. For higher pressures 
the molten salt can be balanced by external 
pressure and used as a zero-reading ma- 

G. Kullerud 

An apparatus for differential thermal 
analysis of volatile or reactive materials 
such as sulfides, selenides, and arsenides 
has been assembled during this past year 
and is now in routine operation. The ap- 
paratus is very similar to that described by 
Kracek (1946) and utilizes many of the 
original components. The present appa- 
ratus can be operated without difficulty 
at pressures exceeding 60 bars with tem- 
peratures up to 1050° C. By using a hori- 
zontal furnace arrangement it has been 
possible to reduce the temperature gradient 
to less than 1° C, over the 2-cm length of 
the silica sample containers. So far this 
apparatus has been used to study parts of 
a number of systems including Cu-S, 
Ni-S, Fe-S, Fe-Se, Ni-As, Fe-As, Ni- 
As-S, Cu-Fe-S, and Fe-As-S. 

The incongruent melting of covellite 
(CuS) at 507° ±3° C (Kullerud, Year 
Book 56, pp. 195-197) has been confirmed 
by heating CuS with excess sulfur and 
recording the heating curve. In the system 
Ni-S the existence of an unquenchable 
high-temperature Ni 3 ±^S 2 phase was veri- 
fied. The stability field of this phase has 
been outlined quite distinctly by this 
method. The melting relations of the 
Nii-aS solid solution series and of vaesite 
have been determined as well as the 
solidus and liquidus curves between these 
compositions. The presence of a field con- 
taining two liquids in addition to vapor 
existing at least over the composition range 
of 45 to 20 weight per cent Ni has been 
strongly indicated by the constancy in 


melting temperatures obtained over this 

The incongruent melting of FeS 2 at 
743° ±3° C (Kullerud and Yoder, Year 
Book 56, p. 187) was recorded by differen- 
tial thermal analysis at 742° ±3° C. 

In the Fe-Se system, FeSe 2 (ferroselite) 
when heated with excess selenium was 
found to melt incongruently at 570° ± 

bergite) ; a eutectic was found to exist at 
852° ±3° C (see Yund, fig. 42). 

Of the ternary compounds NiAsS (gers- 
dorfHte) was found to melt at 890° ±5° C, 
and Cu 5 FeS4 (bornite) was found to un- 
dergo a high-low inversion at about 190° C. 

The D. T. A. apparatus serves as a use- 
ful tool in the areas that can also be in- 
vestigated by means of quench-type meth- 

590 600 770 

Temperature, °C 

Fig. 49. Differential heating and cooling curves on FeSe 2 + Se starting material, showing invari- 
ant conditions at 570° ± 10° C (Fe 8 _ a .Se 4 + FeSe, +L+ V) and at 790° C (Fej. ^Se + Fe 3 _ a .Se 4 + 
L + V). 

10° C to produce Fe 3 Se4 4-L. The FesSe 4 
phase in turn melts incongruently at 789° 
±3° C to Fei-xSe + L. The heating and 
cooling curves registering these reactions 
are shown in figure 49. The melting rela- 
tions of the extensive Fei-^Se solid solution 
series have so far not been studied. 

In the Ni-As system the differential 
thermal analysis apparatus was used to 
determine the melting relations between 
NiAs (niccolite) and NiAs 2 (rammels- 

ods. Its greatest usefulness, however, lies 
in the areas that cannot be explored by 
means of quench experimentation. The 
D. T. A. experimental method makes it 
possible to determine the presence and sta- 
bility fields of unquenchable phases, such 
as for instance the Ni 3 ±tfS 2 phase men- 
tioned above. Although melting curves of 
sulfide systems usually can be estimated 
only crudely by quenching techniques, the 
melting reactions below 1050° C, in any 


part of the sulfide-type systems, can now 
be investigated routinely and accurately by 
simple D. T. A. experimentation. 

H. L. Barnes 

The mechanisms for the movement and 
deposition of metals during the formation 
of sulfide ore deposits are still an enigma. 
Transport in the vapor phase is improb- 
able for the many deposits exhibiting con- 
vincing evidence of deposition at low tem- 
peratures. Moreover, sulfide minerals are 
known to be insoluble in the unoxidizing 
neutral aqueous solutions described in the 
literature. A preliminary appraisal elimi- 
nates the possibility of transport in the va- 
por phase and also suggests that transport 
in a predominantly aqueous solution is un- 
likely. However, measurements have not 
been previously reported of the solubilities 
of metal sulfides at elevated pressures and 
temperatures in aqueous solutions with a 
composition potentially related to ore trans- 
port. An aqueous solution in equilibrium 
with (and capable of depositing) sulfide 
minerals must contain predominantly re- 
duced sulfur in the form of sulfide, S = , and 
polysulfide, Sx = , ions or molecules (Year 
Book 57). Solutions in the system ZnS- 
H2S-H2O meet this condition. 

An experimental reconnaissance of this 
ternary system has been completed using 
the solubility apparatus described in last 
year's report. Modifications since then in- 
clude the substitution of a stainless-steel 
diaphragm backed with water (similar to 
an aneroid barometer) to isolate the pres- 
sure gauge chemically, and the insertion of 
a titanium electrode in the closure of the 
bomb. The electrode has been used to de- 
tect the surface of liquids on tipping the 
bomb in order to measure their volume and 
to measure roughly the conductivity of 
fluids at each pressure and temperature. 

High-density teflon and the liquids n-oc- 
tane, ft-decane, and silicone oil (Dow 
Corning no. 200), though inert in the sys- 
tem, were found to be highly permeable 

to H 2 S and were unsuitable as chemically 
isolating media in pressure gauges. 

H2S. The deviations from ideality of 
unsaturated H 2 S gas found during the past 
year (fig. 50) exceed those reported by 
Reamer, Sage, and Lacey (1950). The dis- 
crepancies may well be due to the release 
of hydrogen caused by reaction of H2S 
with mercury, which was used as a pres- 
sure transducer by Reamer et al. Com- 
pressibility factors for H2S gas need to be 
redetermined over a greater pressure range 
than is given in figure 50. The large vol- 
ume, over 1.1 liters, of the reaction vessel 
used to determine the curves tends to make 
errors of filling negligible and should give 
accurate results over the entire P-T range 
of the apparatus. 

ZnS-H2S. Zn ++ precipitates from aque- 
ous solutions, in general, by reaction with 
the ion HS~. The precipitate, then, even 
if isolated from the liquid, gradually gives 
off H 2 S gas. This mechanism suggests 
that ZnS might form binary compounds 
of the type ZnS • XH2S, which are meta- 
stable at room temperature and pressure. 
By using a high mole ratio of ZnS to H2S 
in the reaction vessel, the measured pres- 
sure is a sensitive function of the distribu- 
tion of H 2 S between the gas phase and the 
solid ZnS. Combination of less than 5 per 
cent of the H 2 S present with the ZnS 
would be readily detectable; however, no 
binary compound was found over the P-T 
range shown in figure 50. 

ZnS-H 2 0. ZnS tends to fragment and 
form a suspension in liquids even when in 
large crystals. For this reason, much of 
this year's effort has been expended on 
the suppression of suspended material in 
order to extract samples from the reaction 
vessel which are truly representative of 
solubilities. Several modifications of the 
experimental procedure were found to be 
necessary. Coarse, natural sphalerite (ZnS) 
was hand-picked and sieved to eliminate 
fines under 1 mm in maximum dimension. 
The volume of the connection between the 
reaction vessel and the sample tube was re- 


duced from about 3 ml to about 0.3 ml. 
Finally, the suspension was given time to 
settle with the connection to the sampling 
tube above the surface of the liquid phase 
before tilting down the reaction vessel very 
slowly to extract a sample. 
The background level caused by the sus- 

The background due to suspensions was 
found to vary from indetectable to a maxi- 
mum of 1 mg/liter. 

Calculated solubilities, based on a large 
number of assumptions, suggest that the 
solubility of ZnS in pure water is about 10~ 7 
mg/liter. This value is well below the 







100 140 

Temperature, °C 
Fig. 50. Compressibility of dry, unsaturated H 2 S gas. 

pensions was determined by analyses of 
samples of pure water shaken with ZnS 
for varying periods and also of the aque- 
ous phase remaining after an experiment 
of several weeks in the ternary ZnS-H 2 S- 
H 2 0. The analytical sensitivity of the 
polarograph used routinely is 0.1 mg/liter; 
because background from the suspension 
was near this level, colorimetric determina- 
tions were made with dithizone using a 
spectrophotometer to verify the results. 

present background and is not susceptible 
of experimental substantiation. 

ZnS-H 2 S-H 2 0. Data on the binary 
H2S-H2O, by Selleck, Carmichael, and 
Sage (1952), have been recalculated and 
plotted as a P-T projection (fig. 51). Be- 
cause it is apparent that low solubilities of 
ZnS would have very little if any effect on 
this binary, nearly all experiments could 
be made in the ternary system while check- 
ing the binary simultaneously. Therefore, 


the experimental points are saturated in 
ZnS but have been superimposed on the 
binary curves for H2S-H2O interpolated 
from the published data. 

It is evident that the phase boundaries 
involving aqueous liquid, gas, and either 

sorption of both the gaseous and the liquid 
phases from a sample (5.23 ml) extracted 
in the normal manner (Year Book 57) 
into a closed system containing a concen- 
trated solution of NaOH + H 2 2 . The 
sample was then analyzed gravimetrically 

rminations of ZnS solubility in L A 
rminotions of HgS solubility in I_a 

20 40 60 80 100 120 

Temperature, °C 

Fig. 51. A Ptotai - ^ projection of the system ZnS-H 2 S-H 2 0. Phase boundaries and contours of 
H 2 S concentration (moles per liter) in the aqueous liquid (L A ) are interpolated from the data on 
H 2 S-H 2 by Selleck et al. (1952). Experimental points represent determinations under conditions 
of ZnS saturation in the ternary system. G, ga;; L s , sulfide liquid; S s , solid hydrate probably 

a sulfide liquid or solid hydrate are not 
altered within the experimental limits by 
the presence of ZnS. A few points deter- 
mined in the binary system also confirm 
the published data on the position of these 
phase boundaries. 

The solubility of H 2 S in the aqueous 
liquid (La, fig. 51) was determined by 
two methods. The first involved the ab- 

by weighing as BaSCX. The second method 
was based on almost filling the reaction 
chamber with H2O until the volume of gas 
phase was negligible. Knowing P, T, the 
volume of the system, and the weight of 
each of the components permitted accurate 
calculations of the H 2 S concentrations. 

The square shown with each determina- 
tion of the H 2 S concentration (fig. 51) de- 


notes the error in the P-T location of the 
sample. In addition to this error, the 
analytical error using either method is 
±0.02 mole/liter. The experimental points 
for the solubility of H 2 S agree within the 
errors of interpolation for the curves from 
the published data and confirm the pre- 
dicted lack of any effect by ZnS on the 
aqueous liquid-gas part of the binary. 

The P-T location of each of the samples 
analyzed for ZnS is shown on figure 51. 
In the aqueous liquid, the solubility of ZnS 
was found to be less than or equal to the 
background level of 1 mg/liter by both 
polarographic and colorimetric analyses. 
ZnS does not form a stable, soluble com- 
plex ion in the aqueous phase in this 
ternary system. 

The calculated pH in the aqueous liquid 
saturated with H 2 S varies from 4.0 at 25° 
C, 1.0 bar, to about 2.9 at 100° C, 100 bars. 
If the pH were fixed by other components 
in a more complicated system, there is, 
nevertheless, little change in the calculated 
solubility of H 2 S (neglecting deviations of 
activity coefficients from unity) over a 
range from 1 to 7. 

pU 1 7 

H.,S (m/1000 g) 0.1 0.1 
HS- (m/1000 g) 10- 7 0.1 

Total H„S solubility 
(m/1000 g) 




0.1 0.2 1 10 

Up to a p¥{ of 7, there should be little 
change in the solubility of ZnS from that 
in the simple ternary system from reaction 
of ZnS with the predominant ion, H 2 S. 

Samples of both dry and water-saturated 
H 2 S liquid were extracted as normal 
samples and removed in a condensed state 
using a dry ice and acetone bath in order 
to determine the ZnS solubilities. Dry, 
liquid H 2 S is known to be a poor solvent 
for most metallic compounds; consistent 
with the generalization, the solubility of 
ZnS was found to be indetectable polaro- 
graphically at less than 0.1 mg/liter at 
both 20 bars, 25° C, and 36 bars, 49° C. 
However, the solubility of ZnS in water- 

saturated H 2 S liquid is 11 ±2 mg/liter 
(fig. 52) at 50° C, 35 bars. The rate of 
equilibration shown in the figure is slow 
because the low-density sulfide liquid is 
separated (except momentarily during agi- 
tation) from the ZnS by the immiscible 
aqueous solution in which ZnS is prac- 
tically insoluble. The sulfide liquid has 
very low viscosity even in comparison with 
water, which may well explain the lack of 
formation of the troublesome suspension 
found in aqueous fluids. 

Approximate conductivity measurements 
were made on both dry and wet sulfide 
liquids. The rough value for the dry liq- 
uid agrees with the value of the specific 
conductance given by Wilkinson (1931) 
at about 10" 11 ohm -1 . Water, in compari- 
son, has a value of about 10" 8 ohm -1 . The 
conductivity of the wet sulfide liquid was 
also very low and too small for quantita- 
tive measurement with the present tech- 
nique, although saturated in both H 2 and 

It is apparent that this sulfide liquid 
is an aprotic solvent, essentially unionized. 
An analogy between the soluble ion, Zn0 2 = , 
in water and the ion ZnS 2 = in liquid H 2 S 
could not be expected to hold when the 
great difference in the degree of ionization 
of the two solvents is known. 

A neutral to slightly acidic and reduced 
aqueous phase is no longer a potential 
solvent for transporting the metallic sul- 
fides in the processes forming ore deposits. 
The solubilities of ZnS are negligible in 
aqueous solutions with an H 2 S concentra- 
tion even considerably in excess of the 
maximum possible concentrations of sul- 
fur-containing ions and molecules at each 
temperature (Year Book 57). 

The solubility of ZnS in a water-satu- 
rated liquid, immiscible with water, and 
of high sulfur content, may present an al- 
ternative to aqueous transport. Although 
ZnS is insoluble in the dry H 2 S liquid, 
saturation by the addition of about 3 mole 
per cent water causes appreciable solubility. 
For several reasons, this water-saturated 



sulfide fluid could not be the transporting 
fluid. Its critical point (100° C and 90 bars, 
Selleck et al.) indicates an excessively high 
partial pressure of H 2 S for low-tempera- 
ture ore deposits. Furthermore, a water- 
saturated sulfide liquid is much too re- 
ducing to deposit many of the oxide and 
polysulfide phases commonly found in 
assemblages of ore minerals. Nevertheless, 

dissolved sulfur (critical point about 1040° 
C and 117.5 bars) in the water-saturated 
sulfide liquid should raise the temperature 
of the critical point, the oxidation state, and 
possibly the ore solubility simultaneously. 
The solubility of ZnS in this polysulfide 
liquid and the characteristics of this liquid 
need to be determined for evaluation as 
a possible ore solution. 




- 1 

- r / 


^Melting of solid H 2 S-6H 2 

during heating 
I I 1 I 

— f — 

II. ± 2. 

_J ° ° ~! 



Time, hours 



Fig. 52. Rate curve showing solubility of ZnS in H 2 0-saturated sulfide liquid at 50° C and 
35 bars. 


S. P. Clar\, Jr., and G. Kullerud 

The widespread belief that the earth's 
mantle is composed of silicates and the core 
of nickel-iron is largely based on the fairly 
sharp division of meteorites into bodies 
having these two compositions. Informa- 
tion about the pressures and temperatures 
under which the meteorites formed will en- 
able earth models based on meteorites to 
be scrutinized more critically than is pos- 
sible at present. Furthermore, it is evi- 
dently desirable to know as much as pos- 
sible about the origin and history of the 
only available samples of the material in 
interplanetary space. 

Iron and nickel usually make up more 
than 98 per cent of the iron meteorites, and 

two metal phases, kamacite (correspond- 
ing to body-centered-cubic a-iron) and 
taenite (corresponding to face-centered- 
cubic y-iron) are almost invariably pres- 
ent. Taenite contains more nickel than 
kamacite, and the compositions of the two 
phases are explicable in terms of the phase 
diagram of the iron-nickel system, which 
is known at low pressure. The nickel con- 
tent of the taenite has in fact been used 
as a measure of the temperature of forma- 
tion of meteorites, but the effects of pres- 
sure and of other elements on the iron- 
nickel system must be known before such 
estimates can be made with any reliability. 
Pressure is known to lower the tempera- 



ture of the a-y transition in pure iron, but 
quantitative data are lacking for the binary 

Cobalt is the only metallic element, other 
than iron and nickel, that is present in 
amounts greater than 0.1 per cent in iron 
meteorites. Its effect on the iron-nickel sys- 
tem is probably small, since the addition 
of 5 per cent cobalt to pure iron changes 
the temperature of the a-y transition by 
less than 10° C and the cobalt content of 
meteorites usually does not exceed 1 per 
cent. Similar considerations suggest that 
other metallic elements present in minor 
amounts do not affect the iron-nickel sys- 
tem by more than about 5° C. 

Small amounts of nonmetallic elements 
may have a much larger effect, however. 
The average bulk composition of 360 iron 
meteorites, as given in Dana's System of 
Mineralogy (1952, p. 120), shows 0.22 
weight per cent phosphorus, 0.11 per cent 
carbon, and 0.16 per cent sulfur. These 
figures represent averages of analyses of 
widely differing reliability made on rather 
heterogeneous material, and no great sig- 
nificance should be attached to the exact 
figures. They indicate, however, that suf- 
ficient carbon is present to depress the a-y 
transition by 100° C or so, and phosphorus 
enough to raise it by a similar amount. 
Sulfur has practically no effect on the a-y 
transition of pure iron. Considerable in- 
terest has recently centered around the sili- 
con content of iron meteorites; available 
analyses show that it is very low (~0.01 
per cent). 

Even if allowance could be made for the 
effect of nonmetallic elements on the iron- 
nickel system, it is unlikely that a sensi- 
tive thermometer and barometer, based 
only on the compositions of coexisting 
kamacite and taenite, can be developed. 
The nickel content of kamacite does not 
change greatly with temperature and will 
probably prove to be little affected by pres- 
sure as well. A relation between tempera- 
ture and pressure can be obtained from the 
nickel content of the taenite, but the com- 
position of a third phase, coexisting with 

both kamacite and taenite, must be used 
to fix the physical conditions. The iron- 
nickel ratios in the sulfide troilite (FeS), 
the phosphide schreibersite [(Fe,Ni)3P], 
or the carbide cohenite [(Fe,Ni) 3 C] may 
provide the additional information that 
would enable us to fix both the tempera- 
ture and the pressure at which the present 
mineralogy of a given meteorite was 

The System Fe-Ni-S 

A part of the system Fe-Ni-S between 
the FeS-NiS join and the binary system 
Fe-Ni has been investigated with two aims 
in mind. Information on the direction of 
the tie lines connecting coexisting alloys 
and nickeliferous troilites was desired, as 
was knowledge of the melting relations in 
this system. Several iron-nickel alloys 
were prepared by vacuum fusion in an in- 
duction furnace, and charges were made by 
adding sulfur to them. All runs have been 
made in evacuated silica tubes at tem- 
peratures of 1000° C or less. 

Melting relations in the part of the sys- 
tem that is interesting in connection with 
meteorites are shown in figures 53 and 54. 
The two isothermal sections have been 
drawn on the assumption that the solu- 
bility of sulfur in the alloys is negligible, 
and that the troilite coexisting with metal 
or melt has a metal-to-sulfur ratio of 1:1. 
There is reason to believe that this last 
assumption is not strictly true, but quan- 
titative measurements of the deviation 
from stoichiometry are so far not available. 
A small three-phase field bounded by two 
alloys, each containing less than 1 per cent 
nickel, and by a troilite with less than 0.5 
per cent nickel, has been omitted from the 
900° C section. 

Textural evidence was used to locate the 
liquidus. Charges were homogenized by 
holding them at temperatures 70° to 100° 
C above that of the run. The temperature 
was then dropped to the desired value and 
held there for 10 to 15 minutes, and the 
charge was quenched in ice water. Metal 


crystals formed at temperature were much 
coarser than those that crystallized on the 
quench, and could readily be recognized. 
Their high density caused them to concen- 
trate at the bottom of the charge. If the 
Ni/Fe ratio of the charge was 0.2 or 
greater, the liquid yielded a fine-grained 
intergrowth of troilite, pentlandite, and 
metal on quenching. Primary crystals of 
troilite, formed at temperature, could 
readily be distinguished in this range of 
bulk compositions. It was not possible to 
follow to low nickel contents that part of 

Fig. 53. Phase relations in the system Fe- 
Ni-S at 900° C. Vapor pressures are those of the 

the liquidus on which troilite was the pri- 
mary solid phase. The formation of pent- 
landite on quenching also enabled us to 
delineate the three-phase field between tro- 
ilite, alloy, and liquid shown in the section 
at 900° C. 

The temperature of the part of the liq- 
uidus on which metal is the primary solid 
phase is extremely sensitive to the sulfur 
content of the charge. Changing it by 1 
per cent by weight changes the tempera- 
ture of the liquidus by about 100° C. In 
the three-phase region at 900° C the liquid 
is richest in nickel and the troilite contains 
practically none. The appearance of pent- 
landite as a quench product has already 
been mentioned. Scattered grains are 

formed even when liquids containing less 
than 10 per cent nickel are quenched from 
1000° C. This mineral has never been 
reported in meteorites. 

The globular nature of the troilites in 
meteorites has led some observers to con- 
clude that they formed from a melt. In 
this system, however, kamacite and liquid 
are incompatible. The addition of phos- 
phorus can be expected to lower the melt- 
ing point and increase the upper limit of 
stability of kamacite, and these effects may 
remove the incompatibility. The common 
association of troilite and schreibersite in 

1000 'C 

Fig. 54. Phase relations in the system Fe- 
Ni-S at 1000° C. Vapor pressures are those of 
the system. 

meteorites may be significant in this con- 
nection. The addition of phosphorus to 
this system would be an important future 

Subsolidus investigations in this system 
thus far consist of a series of runs at 800° C. 
Alloys containing from 5 to 38.6 weight 
per cent nickel were mixed with about 20 
per cent sulfur in order to produce metal 
and monosulfide in nearly equal amounts. 
Pentlandite was not encountered in any of 
these runs. The small change in cell di- 
mension of the alloys prevented determina- 
tion of their compositions by X-ray dif- 
fraction methods. A measurable change in 
the d(102) value of the sulfides was found, 
but its relation to composition is not simple. 
The results are shown in table 10. The data 



TABLE 10. 

J(102) Variation in Sulfide Phase 
at 800° C 







Fe + Ni 





(Fe,Ni)S + alloy 
(Fe,Ni)S + alloy 
(Fe,Ni)S + alloy 
(Fe,Ni)S + alloy 
(Fe,Ni)S + alloy 
(Fe,Ni)S + alloy 











FeS + Fe 

given there do not represent what is usually 
thought of as a spacing versus composition 

relation, since the d values are given as a 
function of the bulk composition of the 
charge rather than as a function of the iron- 
to-nickel ratio in the sulfide. 

The highest d value given in the table 
was found for the troilite that formed in 
equilibrium with an alloy which originally 
contained 10 per cent nickel. If a curve 
of d value versus bulk composition were 
plotted, it would pass through a maximum 
near this value. This maximum is believed 
to be real and not due to errors in measur- 
ing the X-ray patterns. The same remarks 
apply to the difference between the two 
samples of FeS grown with different 
amounts of sulfur present. We are not 
yet in a position to offer an explanation of 
these observations, but it is clear that the 
last word on FeS has not been spoken. 


(A cooperative program of the Geophysical Laboratory and the Department of 
Terrestrial Magnetism of the Carnegie Institution of Washington) 

G. R. Tilton, G. L. Davis, G. W. Wetheritt* L. T. Aldrich, 8 and Emilie ]ager 

In the early years of the application of 
the various dating methods to granites and 
pegmatites, the most impressive aspect of 
the results was the general agreement be- 
tween ages given by several minerals from 
a particular rock, indicating that the time 
of crystallization of the rock could be ac- 
curately determined. Results of this type 
continue to accumulate, but striking ex- 
amples have been found where each of sev- 
eral minerals records a different event in 
the history of a rock. In addition to the 
differences already found between the age 
of biotite and zircon from metamorphic 
rocks, newer work has shown differences 
between microcline, muscovite, and biotite 
in a single rock assemblage consisting of a 
pegmatite and its enclosing granite or 
gneiss. Although the pegmatite was em- 
placed later than the rock that it cuts, 
the muscovite from the pegmatite, as meas- 
ured, appears to be older than the biotite in 
the enclosing granite or gneiss. 

8 Department of Terrestrial Magnetism. 

It is apparent that coexisting minerals 
differ in their reactions to the intensity of 
conditions of metamorphism. The min- 
erals may be completely or partly recrystal- 
lized, depending on their stability, and 
they may have also lost all or part of the 
radiogenic daughter-products by diffusion. 
A consequence of the observed differences 
is that the age of any one mineral cannot 
be used to establish the time of intrusion of 
a rock. Only when the ages of several dif- 
ferent minerals from the same rock agree 
can it safely be assumed that the time of 
intrusion has been determined. The fact 
that samples of biotite collected from geo- 
logically associated rocks consistently meas- 
ure the same age may only reflect the oc- 
currence of a significant change of condi- 
tions in the metamorphic history of the 
rocks. Examples of this kind will be dis- 
cussed, wherein it is obvious that age meas- 
urements using biotite alone do not es- 
tablish the time of intrusion of the granite 
in which it occurs. 

The program of mineral age determina- 



tion continues to be concerned with the 
problems of the Maryland Piedmont, the 
Blue Ridge region of the southern Appa- 
lachian Mountains, and southern Ontario. 
An investigation of age relationships in the 
European Alps was started during the past 
year. The results indicate that the Rb-Sr 
and K-A methods can be profitably applied 
to Alpine rocks, in spite of their geo- 
logically young age. 

ticular significance in that it indicates the 
existence of crystalline rocks in the area 
1100 million years ago. In this gneiss the 
microcline occurs as lenses or "eyes" of 
nondetrital origin. The age of the micro- 
cline strengthens the suggestion made last 
year from studies of the Baltimore gneiss 
that crystalline rocks existed 1100 million 
years ago in this part of the basement com- 

TABLE 11. Ages from the Washington-Baltimore Area 


Age, mill 

ion years 


Rb 87 
Sr 87 

K 40 

A 40 



Pb 207 

Th 232 

p b 206 

Pb 207 

Pb 206 

p b 208 

Baltimore gneiss 

Phoenix Dome 









Towson Dome 










Woodstock Dome 




Hartley augen-gneiss 




Woodstock granite 








Ellicott City granite 








Guilford granite 




Kensington granite gneiss 








Mineral Ages in the Maryland Piedmont 

In last year's report it was shown that 
the Baltimore gneiss contains minerals of 
two different ages, 1100 and 300 million 
years, and that this pattern of ages might 
be explained by a period of Precambrian 
metamorphism followed by a period of 
metamorphism and intrusion 300 million 
years ago, during Paleozoic time. Addi- 
tional measurements have shown that such 
a simple interpretation is inadequate. 

The results of all the measurements on 
minerals from the Baltimore area are sum- 
marized in tables 11 and 12. The age 
measured for the microcline separated 
from the Hartley augen-gneiss has par- 

Comparison of tables 11 and 12 reveals 
a marked difference between the Rb-Sr 
ages measured for muscovite and micro- 
cline from pegmatites and those deter- 
mined for biotite from the associated gran- 
ites and gneisses. In two places, at Ellicott 
City and Guilford, muscovite from peg- 
matites gives older Rb-Sr ages than bio- 
tite from the granites cut by the pegma- 
tites. The K-A ages are generally con- 
sistent and do not show this discrepancy. 

The widespread occurrence of an age of 
300 million years for biotite leaves little 
doubt that this dates a significant geo- 
logical event. On the other hand table 12 
shows that there is a tendency for the peg- 


matite ages to group at 440 to 450 million 
years. This age is found for both micro- 
cline and muscovite in pegmatites to the 
northeast as well as to the west of Balti- 
more. The consistency indicates that these 
pegmatites were intruded about 450 mil- 
lion years ago. The 450-million-year ages 

TABLE 12. Ages from Pegmati 

tes near 










Rb 87 


Sr 87 

A 40 

Daniels, 10 mi west 

of Baltimore 




Loch Raven 1, 11 mi 

northeast of Balti- 





Loch Raven 2, 11 mi 

northeast of Balti- 





Henry ton, 17 mi west 

of Baltimore 








Ellicott City,f 9 mi 

southwest of Bal- 




Guilford,^ 16 mi 

southwest of Bal- 




Atholton, 16 mi 

southwest of Bal- 




* K-A ages determined by Wasserburg, Petti- 
john, and Lipson (1957). These samples were 
made available for Rb-Sr determinations through 
the courtesy of G. J. Wasserburg. 

t Pegmatite intrudes Ellicott City granite of 
table 11. 

t Pegmatite intrudes Guilford granite of table 

cannot be explained by assuming that 300- 
million-year-old minerals incorporated 
strontium of a single abnormal isotopic 
composition when they formed, because 
the percentage of radiogenic Sr 87 varies by 
more than a factor of 2 in the samples. If 
the incorporated strontium were not of 
one specific isotopic composition it would 
be a remarkable coincidence that all these 

minerals give ages close to 450 million 
years. Likewise, the consistency of the 450- 
million-year ages seems to be too marked 
to be explained by loss of different propor- 
tions of rubidium from 300-million-year- 
old minerals. 

From Baltimore southwest to Ellicott 
City and Guilford, the pegmatite age val- 
ues become progressively younger. The 
pegmatites appear to belong to the same 
swarm as those to the northeast and west 
which have ages of 450 million years. It 
is believed that all the pegmatites were in- 
truded 450 million years ago and that those 
with younger apparent ages have been af- 
fected by a later metamorphism. 

Since the Guilford and Ellicott City 
granites are older than the pegmatites, but 
their biotites are younger, the biotites must 
have lost radiogenic daughter strontium 
under conditions such that muscovite and 
microcline in pegmatites retained it. Some 
support for this reasoning is found in the 
work of Eugster and his colleagues with 
annite (ferrous-iron end member of the 
biotite group) and of Yoder and Eugster 
with muscovite. Their work shows that 
annite is less stable than muscovite under 
some conditions. The stability of a particu- 
lar biotite, however, depends on the partial 
pressure of oxygen and the magnesium/ 
iron ratio. The stabilities of both biotite 
and muscovite vary with surrounding min- 
erals available for reaction. It is accord- 
ingly difficult to say what might have hap- 
pened to the micas in any particular rock. 

It is important to realize that the "age" 
of a mineral may have a number of mean- 
ings. Ideally, it is a measure of the time 
that has elapsed since crystallization from 
a system in which the constituent elements 
were mobile of a mineral that contained no 
radiogenic daughter isotope, or only the 
proportion of that isotope found in the 
normal element. Total or partial loss of 
daughter isotope by diffusion will give an- 
other meaning to the "age." The measured 
age will not show the time of crystalliza- 
tion in this case but will indicate some 


more recent date. All that is required is 
the existence of sufficiently high tempera- 
tures and proper concentration gradients 
to allow diffusion to take place. The 300- 
million-year ages measured for biotite in 
the Ellicott City and Guilford granites 
could indicate this type of loss of daughter 
isotope since it is probable that these gran- 
ites were intruded at least 450 million years 

Knowledge of the diffusion rates of 
strontium and argon in biotite and mus- 
covite is lacking. Gerling and Morozova 
(1957) have measured the activation en- 
ergy for the liberation of argon from a 
muscovite and a biotite. The respective 
values were 85 and 57 kcal/mole, showing 
that argon was more tightly bound to the 
muscovite. This information suggests, but 
does not prove, that the rate of diffusion 
of argon in the biotite is greater than that 
in the muscovite. 

In any event, the data indicate that, al- 
though pegmatites (and probably granites 
as well) were intruded 450 million years 
ago, biotite has retained radiogenic stron- 
tium and argon only for the last 300 mil- 
lion years. For the past 300 million years 
temperatures have evidently been low 
enough so that biotite has lost neither ar- 
gon nor strontium by diffusion. Recent 
work on the diffusion of argon in biotite 
by Everenden, Curtis, Kistler, and Obrado- 
vich (1959) indicates that temperatures of 
300° to 400° C for a period of several mil- 
lion years completely remove argon from 
the mica. They used "books" of mica 200 
microns thick, comparable to the average 
thickness found for biotite in granites. 
This gives some idea of the minimum tem- 
peratures required for argon removal, but 
similar data are lacking for strontium. 
The age data themselves provide no in- 
formation about temperatures in the inter- 
val between 450 and 300 million years, 
except that some 450-million-year pegma- 
tite ages were preserved. 

The 150-million-year interval between 
the two ages is reasonable in the light of 
evidence on crustal movements that have 

taken place in the Appalachians. In New 
England and the Maritime Provinces, 
where fossils are more abundant, episodes 
of crustal movement can be recognized 
from late Ordovician through the Silurian 
and Devonian into the early part of the 
Mississippian period, an interval of 150 to 
250 million years as nearly as can be esti- 
mated from present knowledge of the fossil 
time scale. The movements have been di- 
vided into three orogenies, since three fossil 
horizons are available to distinguish them, 
although movements were quite certainly 
much more continuous than this rather 
arbitrary division suggests (King, 1951, 
pp. 78-79). The orogenies in the northern 
Appalachians have not been definitely 
traced into the nonfossiliferous rocks of the 
Baltimore area. It is possible, however, 
that the 450-million-year ages are due to 
crystallization of granitic rocks attending 
one orogeny whereas the 300-million-year 
ages are the result of less intense meta- 
morphism accompanying a later orogeny. 
If crustal movements were more or less 
continuous, the 300-million-year ages 
would represent a late stage of a single 
long orogenic episode. 

Muscovite in the Guilford granite has a 
Rb-Sr age younger than that of the musco- 
vite in the pegmatite. A similar observation 
has been made at Cutler, Ontario (see 
table 14). This may be a reflection of the 
smaller grain size of the muscovite in the 
granite or of compositional differences in 
the two rocks. The granites have some- 
what discordant zircon ages, but the Pb 20T - 
Pb 206 ages agree reasonably well with the 
older pegmatite ages. The U-Pb and Th- 
Pb age values are similar, but none are 
less than 300 million years, the prevailing 
age of the biotites of the area. 

To summarize: there is clear evidence of 
a Precambrian metamorphism 1100 million 
years ago. The Baltimore gneiss and Hart- 
ley augen-gneiss were crystalline by that 
time. It is believed that in the Paleozoic 
era (450 million years ago) there was in- 
trusion of pegmatites (and probably gran- 
ites). Age measurements on biotite indi- 


cate still another event 300 million years 
ago whose geologic significance is less 
clear. Perhaps it is to be correlated with a 
later metamorphic episode that was less 
intense than the one 450 million years ago. 
In New England and the Maritime Prov- 
inces, orogenies are known to have oc- 
curred throughout a time interval of com- 
parable span. 

granites and a rare-earth vein that intrude 
the gneiss. The gneiss at Mortimer, N. C, 
has been called Cranberry in reconnais- 
sance studies by Keith, but it is separated 
from true Cranberry gneiss by faults; 
hence this identification is uncertain. 

The mica ages in table 13 follow no 
fixed pattern. The Rb-Sr age of 350 mil- 
lion years at Deyton Bend agrees rather 

TABLE 13. Southern Appalachian Ages 



Age, mi 

llion years 


Rb 87 
Sr 87 

K 40 

A 40 

p b 206 


Pb 207 

^h 232 

Pb 207 


Pb 208 

Deyton Bend, N. C. 

Cranberry gneiss 








Pardee Point, Tenn. 

Cranberry gneiss 








1 mi southeast of No- 

where Ridge, Tenn. 

Cranberry gneiss 




Mortimer, N. C. 







Roan Mountain, Tenn. 

Beech granite 








Crossnore, N. C. 







Laurel Gap, Tenn. 

Rare-earth vein 






Shenandoah National 

Park, Va. 









Spruce Pine, N. C. 







Mineral Ages in the Southern 

Earlier work has established the pres- 
ence of minerals with ages of 1100 million 
years from New York to Virginia in the 
Appalachian orogenic belt. In the past 
year considerable work has been done in 
an effort to extend the known range of oc- 
currence of these old minerals to North 
Carolina and Tennessee. The results are 
shown in table 13, together with previous 
results from a pegmatite at Spruce Pine. 
All samples were collected in the Blue 
Ridge Province, where the basement com- 
plex is exposed. The members of the base- 
ment studied were the Cranberry gneiss of 
presumed sedimentary origin, and two 

well with the age of 375 million years es- 
tablished for pegmatites in the Spruce Pine 
district. The age of the biotite from Pardee 
Point agrees with that at Shenandoah Na- 
tional Park, where an age of 1100 million 
years has been established for zircon from 
the gneiss. Long and Kulp at the Lamont 
Geological Observatory found a 900-mil- 
lion-year age for the biotite at Pardee Point 
but did not measure zircon ages. The re- 
maining mica values are probably not true 
ages but instead represent Precambrian 
micas that lost variable amounts of stron- 
tium and argon during the Paleozoic era. 
All the ages of zircons from North 
Carolina and Tennessee are discordant. 
This regional occurrence of discordant ages 


is interesting in itself and contrasts sharply 
with the nearly concordant ages found for 
zircon in Precambrian rocks studied in 
Virginia, Maryland, and New York. The 
Pb 207 -Pb 206 age of 1270 million years at 
Deyton Bend records a history that goes 
back even farther than the 1100-million- 
year ages found elsewhere. Since the zir- 
con occurs in a rock of probable sedi- 
mentary origin, this result cannot be used 
to prove or disprove the existence of crys- 
talline rocks in the area earlier than 1100 
million years ago. 

A study of the discordant ages for the 
Beech granite and the Mortimer, Cross- 
nore, and Laurel Gap zircons leads to an 
interesting conclusion. A graphical method 
of analyzing the discordant ages is shown 
in figure 55. The method has been dis- 
cussed fully in the report of the Director 
of the Department of Terrestrial Mag- 
netism for the year 1955-1956, but a brief 
description will be given here. If a min- 
eral has a U 238 -Pb 206 age equal to its U 235 - 
Pb 20T age, the ages are said to be con- 
cordant. The concordant age of a mineral 
can be represented by a point on a plot of 

p b 206/TJ238 against p b 207/TJ23 5i ^ ^ 

points determine a concordant age curve 
called "concordia," along which a time 
scale may be marked off. Thus a mineral 
having an age of 1000 million years and 
having lost neither uranium nor lead has 
Pb 206 /U 238 and Pb 207 /U 235 ratios that will 
lie on the curve at 1000 million years. Next 
consider the effect of loss of lead 375 mil- 
lion years ago from a group of minerals 
having an age of 1000 million years. Min- 
erals that lost all their lead will plot on 
concordia at 375 million years. It can be 
shown that minerals that lost only part of 
their lead will plot on the chord connect- 
ing 1000 and 375 million years. Their posi- 
tion on the chord will be determined by 
the fraction of lead lost. The four samples 
from the southern Appalachians fit the 
chord almost within experimental errors, 
suggesting that all these zircons are 1000 
to 1100 million years old. The analysis indi- 
cates that Paleozoic metamorphism in the 

area 350 million to 400 million years ago 
can account for the discordant ages and 
that these zircons have not gained or lost 
uranium and lead since that time. 

When the zircon evidence in figure 55 is 
taken together with the 900-million-year 
Rb-Sr age at Pardee Point, it appears that 
the 1000-million- to 1100-million-year meta- 
morphism found from New York to Vir- 
ginia does extend farther south into North 
Carolina and Tennessee. 

Fig. 55. Lead-loss pattern for zircons from 
the basement of the southern Appalachians. The 
chord is the locus of points representing minerals 
1000 million years old which lost lead 375 mil- 
lion years ago. 

That the Rb-Sr ages for biotite in table 
13 measure 900 million years, but not any 
older, is significant in view of the fact that 
biotites at Bear Mountain, N. Y., and the 
New Jersey highlands also give Rb-Sr ages 
that equal but do not exceed 900 million 
years. The time interval between the 900- 
million-year biotite ages and the 1100-mil- 
lion-year zircon and feldspar ages is analo- 
gous to the interval between the 300-mil- 
lion-year biotite ages and the 450-million- 
year pegmatite ages in the Maryland Pied- 
mont. It is possible that the 1100-million- 
year-old rocks experienced a degree of 
metamorphism 900 million years ago that 
removed radiogenic strontium from biotite 


but did not seriously affect zircon as at 
Shenandoah National Park. 

Comparison of Ages of Minerals from 

Pegmatites and Enclosing Granite 

at Cutler, Ontario 

The discussion of the Maryland Pied- 
mont has presented the age relationships 
measured between muscovite and micro- 
cline in pegmatites and biotite in the rocks 

TABLE 14. Ages from the Cutler Batholith, 
Cutler, Ontario 






Rb 87 
Sr 87 

K 40 

A 40 

Pegmatite 1 







Schist cut by 

P e g- 

matite 1 




Pegmatite 2 




Granite A cut by 









Granite B 







Granite C 




* Zircon from granite A gave the following 

million years 

TJ238/p b 206 730 

TJ235/p b 207 910 

p b 207/p b 206 147 

Th 232 /Pb 208 94 

cut by and therefore older than the peg- 
matites. At Cutler, Ontario, a more com- 
plete study has been made where similar 
relationships exist; table 14 summarizes 
the results. Muscovite and microcline from 
the pegmatites give consistent Rb-Sr ages 
of about 1750 million years. Biotite from 
granites and schist gives concordant Rb-Sr 
and K-A ages of 1300 to 1400 million years. 
By Rb-Sr measurement, the muscovite 
from the granite is appreciably older than 
the biotite from the same granite, but not 
as old as the muscovite from the pegma- 

tites. This age pattern is the same as the 
one for the Maryland Piedmont. The meas- 
urements indicate that the pegmatites at 
Cutler were intruded 1750 million years 
ago. Later, about 1300 million years ago, 
the rocks were subjected to conditions of 
metamorphism sufficiently intense to al- 
ter the K-A age of muscovite in the peg- 
matite and the K-A and Rb-Sr ages of 
micas in the host rock. Zircon from one 
of the granites has discordant ages that 
are of no apparent help in interpreting the 

Other examples of this age pattern have 
been found. Table 15 gives the results of 

TABLE 15. Ages from a Pegmatite and Enclos- 
ing Gneiss near Randville, Michigan 



million years 

Rb 87 
Sr 87 

K 40 

A 40 

Gneiss * 




* Zircon from the gneiss gave the following 

million years 

TJ238/p b 20C 1710 

TJ235/p b 207 2100 

p b 207/p b 20G 2500 

age measurements of minerals from a 
gneiss-pegmatite assemblage in northern 
Michigan. The K-A age of biotite in the 
host rock shows the effects of metamor- 
phism at the same 1300-million-year date as 
at Cutler, but the K-A age of the muscovite 
and the Rb-Sr age of the biotite suggest 
that metamorphic intensity was lower at 
Randville. The effects of metamorphism 
on mineral ages in northern Michigan are 
discussed more completely in the report of 
the Director of the Department of Ter- 
restrial Magnetism. Additional examples 
of pegmatite minerals that have older ages 
than minerals from rocks cut by the peg- 
matites have been reported at Death Val- 


ley, California, by Wasserburg and Wether- 
ill and in Karelia by E. K. Gerling. 

Preliminary Study of the Age Relation- 
ships of Some Alpine Roc\s 

The Alpine orogeny during the Tertiary 
period, about 10 to 60 million years ago, is 
one of the most recent major upheavals in 
the earth's crust. During the orogeny, 
large sheets of deeply buried rocks were 
thrust up and over less deeply buried rocks 
to form nappes. These rocks appear to have 
been recrystallized in varying degree as a 
result of the orogeny. The Alps thus pro- 
vide the basis for a study, similar to that 
already under way in the Appalachians, 
in which an attempt can be made to date 
the orogeny and to determine its effect on 
the ages of minerals in rocks already pres- 
ent. Studies in the Alps have the ad- 
vantage of dealing with an area that has 
been mapped in great detail with con- 
siderable geological control available from 
fossil evidence. Because the orogeny is re- 
cent, it can only be dated if minerals 
formed during the orogeny contain very 
small quantities of primary lead, stron- 
tium, or argon; otherwise the small 
amount of radiogenic daughter isotope 
would not be determinable. 

The rocks investigated are a granite, a 
somewhat metamorphosed granite, and a 
gneiss. The results are given in table 16. 
The gneiss from Brione is representative 
of the lower Penninic nappes. It has been 
strongly metamorphosed, petrofabric stud- 
ies indicating complete recrystallization of 
the micas during the main phase of Alpine 
folding in this region. The age of about 
17 million years may date the actual time 
of thrusting. Unfortunately, the age can- 
not be used to better define the strati- 
graphic (fossil) time scale because the time 
of thrusting cannot be accurately related 
to that time scale. Two Rb-Sr age deter- 
minations were made on the biotite from 
this rock, each giving an age of 17 million 
years. The Sr 87 was 8 per cent radiogenic 
in one case and 12 per cent radiogenic in 

the other, so that the age is not highly 
sensitive to the common strontium correc- 
tion. The K-A age of muscovite from the 
gneiss as determined in collaboration with 
H. Faul gave a result of 19 million years. 
The granite from the Aar massif experi- 
enced relatively mild metamorphism dur- 
ing the Alpine orogeny. It is generally re- 
garded as a remnant of a Hercynian range 
(of late Paleozoic age, about 350 million 
years old) that was thrust up in the 
orogeny. Pre-Alpine texture is largely pre- 
served in the rock. Biotite occurs as bent 
or torn grains, some having the usual 
brown color and others, younger appear- 

TABLE 16. Ages from Alpine Rocks 




Location and 
Geologic Unit 



Rb 87 

K i0 

A 40 

Brione, Ticino 

Penninic nappes 






Aar massif 





Mont Orfano 

Southern Alps 







* Determined in collaboration with H. Faul at the 
U. S. Geological Survey. 

ing, being olive green. The sample ana- 
lyzed was a mixture of both kinds of bio- 
tite, and the age value of 75 million years 
can be interpreted as being due to a mix- 
ture of micas having two different ages. 
In the absence of further evidence, it could 
also indicate complete recrystallization 75 
million years ago. 

At Mont Orfano, Italy, 45 km southwest 
of Brione, the unmetamorphosed Baveno 
granite occurs south of a zone of basic 
rocks, metagabbros, and amphibolites 
called the Ivrea zone. This zone has been 
shown, by contact relationships, to be older 
than the granite. The granite itself has 
variously been considered to be as young as 
Alpine or as old as Hercynian. The age 


measured for the biotite as well as the mi- 
crocline is 290 million years, lending sup- 
port to the older date proposed. The Ivrea 
zone is therefore at least late Paleozoic in 
age and could not have crystallized during 
the Alpine orogeny. 

This preliminary work indicates that in 
favorable cases the young age of the Alps 
will not prevent the application of the 
Rb-Sr and K-A methods to rocks of Alpine 
age. Further studies will be made at the 
University of Bern. 


Bruce Bryant and John C. Reed, Jr., of 
the U. S. Geological Survey, supervised the 
collection of many specimens from North 
Carolina and Tennessee. Further advice 
was obtained from Stuart W. Maher, of the 
Tennessee Department of Conservation. 
We have continued to benefit in our work 
in Maryland from discussion with the 
members of the Department of Geology of 
The Johns Hopkins University, particu- 
larly C. A. Hopson. 


W. F. Libby 

The intensive series of bomb tests fired 
by the Soviet Union during October 1958 
affords a unique opportunity to test 
whether stratospheric radioactive fallout 
from injections made at polar latitudes dif- 
fers appreciably in distribution or fallout 
rates from that due to equatorial explo- 
sions, such as the United States and United 
Kingdom have fired. The Russian October 
series is estimated, on the basis of assump- 
tions previously described (Libby, 1958), 
to have added about 12.5 to 15 megatons 9 
equivalent of fission products to the strato- 
sphere, whereas the previous inventory on 
the same basis was about 18 megatons 
equivalent, if we anticipate the result, dem- 
onstrated later, that polar debris falls out 
in about 1 year instead of 5. The Russian 
addition amounted to a sudden increase of 
about 150 per cent in the northern hemi- 
sphere, if we take the previous strato- 
spheric burden to have been uniformly 
distributed. In line with the suggestion of 
Martell (1959) that nuclear explosions that 
are conducted in polar latitudes and inject 
radioactive fallout into the stratosphere 
may have a stratospheric residence time of 
about 1 year, a time much shorter than 
for tests conducted in equatorial latitudes, 

9 It is useful to note that 1 megaton fission 
spread uniformly world wide gives 1/2 mc 
Sr 90 /mi 2 or, if restricted to one hemisphere and 
spread uniformly, gives 1 mc/mi 2 . 

which have a residence time of the order of 
several years (Libby, 1958) , the Department 
of Defense (Shelton, 1959) has tentatively 
concluded on the basis of data that it has 
collected, together with the Atomic Energy 
Commission's stratospheric balloon data, 
that the residence time for polar shots is 
about 1 year or less and that equatorial 
debris has a residence time of about 3 years. 
L. Machta (1957, 1958, 1959) has empha- 
sized that the well established nonuni- 
formity of the total fallout in the northern 
hemisphere with a peak in the middle lati- 
tudes might be due to a natural character- 
istic of the circulation of the stratosphere 
which concentrates stratospheric fallout in 
these latitudes. Libby (1958) has pointed 
out that major test sites in the United 
States and the Soviet Union lie in these lati- 
tudes. He suggests that, because it is air- 
borne only a month or so on the average 
and consequently has relatively little 
chance to spread in the north-south direc- 
tion, tropospheric fallout from these sites 
could contribute a band of early fallout 
around the earth in the general latitude of 
the test site, and that this might account 
for a large part of the peak noted by 
Machta. Comparison of foreign with do- 
mestic soil Sr 90 data demonstrates that the 
United States has about 15 mc/mi 2 more 
Sr 90 than foreign countries in the same lati- 
tudes. Therefore a part of the United 



States fallout was tropospheric and specifi- 
cally due to tests in Nevada. The data con- 
tinue to raise the old question whether this 
excess of the northern over the southern 
hemisphere could all be due to tropospheric 
debris or might be due in one way or an- 
other to stratospheric fallout. 

Study of the soil data, together with re- 
cent rain data, indicates that the true situa- 
tion probably lies in a combination of Dr. 
Martell's theory, Dr. Machta's theory, and 
a part of the author's model (Libby, 1956a, 
1956*, 1957). 

rainfall sample. This measurement can be 
made quickly. In this way the data in 
figure 57 for rain collected at the Geo- 
physical Laboratory in Washington, D. C, 
have been obtained and are given in detail 
in table 17. They show an average fallout 
rate over the last 5 months of about 2.3 mil- 
licuries of Sr 90 per square mile per month, 
a rate that can only be explained by the 
Russian October fallout coming down with 
about a 1-year residence time as Martell 
has suggested. Data from Bedford, Mass., 
Pittsburgh, Pa., and Westwood, N. J., are 
























^^» ^ ~> 5I D * y DECAY CURVES 








1 ^1 ^S. 1 ^^ 






Fig. 56. Spring rain, Sr 89 /Sr 90 ratio. 

In order to observe the fallout from the 
Russian October series promptly and with 
adequate accuracy, the standard procedure 
of Sr 90 fallout analysis was changed to al- 
low up-to-date analyses of rainfall occur- 
ring during the extremely crucial periods 
of March, April, and May 1959. The 
change consisted in a reliance on the 
smoothness of the curve of the ratio of 
Sr 89 /Sr 90 (fig. 56) versus time, in view of 
the fact that no nuclear explosions have 
been fired recently, and the use of the 
knowledge of this ratio by extrapolation of 
the curve to allow one to calculate the Sr 90 
fallout from a measurement of the total 
radioactive strontium (Sr 89 + Sr 90 ) in a 

included in the figure to show the general 
agreement. The mean residence time is cal- 
culated, as shown in figure 57, by subtract- 
ing the 1958 rate for Pittsburgh after 
removing the April 1958 rise which it 
showed, presumably due to the Russian 
tests at the end of February 1958, multiply- 
ing by the ratio of the mean annual rain- 
fall to that for Washington. Thus we see 
that there is a difference in residence time 
for polar and equatorial shots injecting into 
the stratosphere, and we consequently are 
led to speculate about its cause and whether 
the residence time for intermediate lati- 
tudes would be intermediate between 1 and 
5 years. 







1.8 me/ml 2 / mo 



Fig. 57. Spring 1959 fallout from Russian October series. Data from Bedford, Mass., personal 
communication from Dr. E. A. Martell, Air Force Cambridge Research Center; data from Pittsburgh, 
Pa., personal communication from Nuclear Science and Engineering Corp.; data from Westwood, 
N. J., personal communication from Isotopes, Inc. 


TABLE 17. Summary of Washington-Russian October Data 

nt Rainfall, Sr 89 , dpm/1 . ,. „ 89/s 90 Sr 90 Rate, Incremental Total Sr 99 , 

Date in. on date Sr ' dpm/1 bl /bl uc/mi 2 /in. Sr 90 , uc/mi 2 mc/mi 2 

Jan. 1 0.88 203+ 1.7 7.4± 0.4 27.6±1.7 222± 14 196± 11 0.20 

Jan. 14-15 0.12 890± 9 28.6± 1.3 31.2±2 860+ 37 104± 5 0.30 

Jan. 15-16 0.34 2230± 65 48.6± 9.3 45.9±9 1460+280 496± 96 0.80 

Jan. 20 0.33 639± 55 (31)* (202) 1.00 

Jan. 21-22 0.42 (31) (202) 1.20 

Jan. 26-27 0.76 1450± 13 45 ± 2 31.8±1.9 1365± 60 1040± 46 2.24 

Feb. 3^1 0.43 259± 3 10.1± 1.1 25.7±3 303± 33 131 + 14 2.37 

Feb. 12-14 0.75 722± 15 27.1± 1.7 26.7+2 815± 51 611± 28 2.98 

Feb. 23 0.18 1085+ 40 80 ±14 13.9+3 2410+420 434+ 76 3.42 

Mar. 4-5 Trace 950+ 40 42+4 22.6+2.4 1260+126 

Mar. 5-6 1.13 724+203 32.4+ 2.6 23 +2 974+ 78 1104+ 88 4.52 

Mar. 11-12 0.52 446+ 25 35 +13 13 +5 1050+390 544+203 5.06 

Mar. 27 0.44 580+ 30 (13) (585) 5.65 

Mar. 30 0.30 636+ 36 (13) (436) 6.09 
Apr. 2 "1 

6 A.M. to H).29 276+ 50 23.4+ 2 15.3±1.3 700+ 60 519+ 44 6.60 

1 P.M. J 
Apr. 2 ] 

1 P.M. to f0.45 397+ 17 

5 P.M. J 

Apr. 3-4 0.69 236+ 16 (13) (378) 6.98 

Apr. 9-10 0.31 550+ 50 51.5± 3.7 10 +1 1545+110 479+ 35 7.46 

Apr. 10-12 1.17 232+ 13 (12) (688) 8.15 

Apr. 19-20 0.12 1337+ 27 (11) (440) 8.59 

Apr. 27-28 0.65 553± 55 (10) (1000) 9.59 

Apr. 28-29 0.48 380± 20 50.5±15 7.6+2.3 730+216 10.32 

May 2-3 0.11 (150) 10.47 

May 12-13 0.12 (180) 10.65 

May 13 1.05 120+ 5 (8) (430) 11.08 

May 14 0.015 ... ... ... ... ... 11.20 

May 23 1.34 52+ 3 14.9+ 1.5 3.5+0.4 447+ 46 596+ 60 11.80 

May 31 0.06 ... ... ... ... ... 11.88 

June 1-2 2.12 51+2 ... ... ... (577) 12.46 

June 12 0.69 70+6 ... ... ... (390) 12.85 

* Numbers in parentheses are derived by calculating, from the curve as drawn in figure 57, the ex- 
pected Sr 89 /Sr°° ratio and using the observed total strontium activity in the sample of rainfall in- 
volved to calculate the Sr 90 contribution, and thus to calculate the total fallout in the storm in 


P. H. Abclson 

Kerogen makes up about 90 per cent of stances, many of high molecular weight. 
the organic matter in sedimentary rocks By pyrolysis, reduction, and other means it 
and in total quantity in this country can be broken down into smaller mole- 
amounts to more than 1000 times that of cules, including asphaltenes, some of which 
our petroleum reserve. From the kerogen are in the range of 5000 in weight. Kero- 
of the Green River Shale alone about 10 12 gen has the insolubility characteristics of 
barrels of oil could be recovered by retort- many very large polymers, being insoluble 
ing, or about 30 times as much as our in alkali, hydrochloric acid, and organic 
proved petroleum reserves. solvents. It reacts readily with oxygen, and 

Kerogen is a complex mixture of sub- finely divided samples may ignite in air at 


100° C. Associated with it in sedimentary 
rocks are asphaltenes (~10 per cent), hy- 
drocarbons ("-T per cent), and some other 
relatively small molecules such as porphy- 

Because of its inertness to solvents, kero- 
gen is a difficult material to study. The 
ultimate analyses of many kerogens are 
known. These contain variable proportions 
of carbon, hydrogen, nitrogen, oxygen, and 
sulfur, and a typical analysis would be C, 
60; H, 7; N, 2; O, 30; S, 1. At elevated 
temperatures kerogen decomposes to a vast 
variety of smaller molecules whose rela- 
tionship to the parent material is not clear. 
Alternative methods for breaking up kero- 
gen or rendering it soluble are essential to 
progress in studying the material. 

Another possible source of knowledge 
about kerogen is a more detailed under- 
standing of mechanisms of its formation. 
Here also, however, there is an impasse. 
All authorities agree that biosynthetic ac- 
tivities of living matter are the ultimate 
source of kerogen. But the familiar lipides, 
carbohydrates, and proteins of the intact 
algae or the corresponding constituents of 
the microorganisms of the anaerobic sedi- 
ments are obviously not very similar to 
kerogen chemically. True, the ultimate 
analysis shows living matter and young 
kerogen to be somewhat alike. Chemical 
behavior of the two is utterly different, 
however, and kerogen is not a direct meta- 
bolic product. Thus we see there are two 
important and difficult problems concern- 
ing this complex material — its mode of for- 
mation and its structure. Work at this 
Laboratory has met some success in attack- 
ing both questions. 

Mechanisms of Formation of Kerogen 

The fate of fixed nitrogen in sediments 
furnishes an important clue to the forma- 
tion of kerogen. Almost all nitrogen em- 
ployed by living matter is in the form of 
proteins, peptides, and their amino acids, 
with lesser amounts present as nitrogenous 
bases in nucleic acids, nucleotides, or their 

constituent purines and pyrimidines. These 
nitrogen-containing materials are soluble 
in acid and can easily be isolated from liv- 
ing matter. However, in only a short time 
after deposition most of the nitrogen is 
found bound in organic combinations that 
are insoluble. These compounds do not ap- 
pear to be products of enzymic action. Bio- 
logical attack leads to soluble products, 
such as ammonia and amines. Further- 
more, the amino acids are sufficiently stable 
to withstand thermal degradation. The 
major remaining mechanism to account 
for the change in the behavior of nitroge- 
nous compounds is chemical combination. 

Search of the literature reveals that food 
technologists have been concerned with a 
closely related problem — the nonenzymic 
browning of foods. Foods, of course, arise 
from biologic activity and contain varying 
proportions of the familiar major compo- 
nents — lipides, carbohydrates, and proteins. 
On storage many foods undergo chemical 
changes, a noticeable result of which is a 
darkening in color. Since this phenomenon 
is of great economic importance, a number 
of studies have been devoted to it. Many 
of them have involved laboratory investiga- 
tions of model systems, including the reac- 
tions of simple carbohydrates with amino 
acids, peptides, and proteins. The subject 
has been reviewed by Hodge (1953). Only 
part of the processes is understood. One set 
of observations, though, seems particularly 
applicable to the nitrogen and kerogen 

When a typical carbohydrate, glucose, 
and an amino acid, glycine, react, the first 
step is a reversible condensation of the 
amine and aldehyde groups. This is fol- 
lowed by a slower and irreversible reaction 
called the Amadori rearrangement which 
changes the 1-substituted glucose into a 
1-substituted deoxyfructose. In its behavior 
toward acid hydrolysis this last compound 
is reminiscent of nitrogenous combinations 
in kerogen. It is very reactive. It tends to- 
ward polymerization to form complex 
highly colored larger molecules. It is also 



a more highly reducing substance than glu- 
cose and combines with oxygen at room 

Most of the studies made by others on 
the reaction of glucose and glycine have 
been carried out at relatively high or low 
p¥l. The range of interest geologically is in 
the vicinity of pW 8. Hence, a series of ex- 
periments were made using 1 molar glucose 
and 0.1 molar glycine with various buffers. 
These were incubated for 90 minutes at 
100° C, then tested with ninhydrin. Re- 
sults are shown in table 18. Ninhydrin 
reacts both with free amino acid and with 
glycosyl glycine to produce a color reaction. 
The low values shown in table 18 obtained 

TABLE 18. pH Dependence of Reaction of 
Glucose and Glycine 



Left, % 
















Incubated at 100° C for 90 minutes. 

at pYL 7 and higher are indicative that these 
two compounds have disappeared, prob- 
ably as a result of the combining of glucose 
and glycine followed by the Amadori re- 

Samples of a solution containing 1 molar 
glucose and 0.02 molar glycine at /?H 10 
were incubated at 100°, 68°, 51°, and 35° C. 
The time required for ninhydrin values to 
drop to 36 per cent was, respectively, 0.8 
hour, 16, 96, and 400 hours. Extrapolation 
to lower temperatures indicates that at 
20° C the corresponding time would be 
about 70 days. The rate of conversion of 
amino acids in nature is not known pre- 
cisely but such observations as have been 
made are in essential agreement with labo- 
ratory findings. 

All the alkaline reaction mixtures used 
in table 18 developed dark brown colors, a 
phenomenon noted by others. The inten- 
sity of the color is related to the concentra- 
tions of the reactants and the duration of 
incubation. Using a solution at p¥l 8 con- 
taining 2 molar glucose and glycine, very 
deep colors were observed after a day at 
100° C. At 500 mu the optical density of 
these solutions amounted to 600, with 
much higher values at shorter wavelengths. 

Results illustrating the effect of glycine 
concentration and time are shown in 
table 19. A glucose solution at pH 8 was 
incubated with various amounts of glycine 

Incubation at lower temperatures also 
produced colored solutions. A mixture of 

TABLE 19. Optical Densities of Reaction Prod- 
ucts Glucose + Glycine 


Glucose Glycine 

Optical Densities 
at 500 mu 

80min 18 hr 

2 molar 



2 molar 






2 molar 






2 molar glycine and glucose atpH 8 turned 
brown on standing at 4° C for 3 days. 

One of the effects attending these reac- 
tions is formation of substances of greater 
molecular weight. The polymerization of 
glucose at low pYL and high temperature 
has been studied by Mora and Wood 
(1958). By removing water, which is a 
product of the reaction, they were able to 
make polymers containing more than 150 
units of glucose. This polymerization proc- 
ess has not been investigated at geologically 
interesting values of pW. Some results 
obtained this year indicate that larger 
molecules are indeed formed even in the 
presence of excess water at pH 8.0. Thus, 
solutions 2 molar in glucose and glycine 
incubated a day at 100° C yielded sub- 
stances having a molecular weight of about 


These materials represent a substantial 
step from glucose and glycine toward as- 
phaltene or kerogen. Thus, they are inter- 
mediate in molecular weight, highly col- 
ored, and reactive with oxygen, and their 
fixed nitrogen does not respond to ninhy- 
drin. The synthetic products differ from 
asphaltene and kerogen in a major respect, 
however — solubility. These new entities 
are readily soluble in water but not in 
organic solvents. This difference would 
tend to disappear if internal rearrange- 
ments occurred involving elimination of 
water and formation of carbon-carbon 
double bonds. 

The preceding discussion has been based 
on results from a very simple model sys- 
tem. In living organisms sugars are pres- 
ent in a variety of forms — glycogen, for in- 
stance, which is a polymer of glucose. 
Amino acids are present as proteins. Autol- 
ysis following death results in some split- 
ting of these molecules, but the actual re- 
acting components would in large degree 
probably be carbohydrate polymers and 

The existence of similarities between the 
simple model systems and asphaltene and 
kerogen does not necessarily prove that 
knowledge derived from the simple labo- 
ratory tests is directly applicable to events 
in the sedimentary rocks. The leads are 
promising, however, and it appears likely 
that much of the early course of events in 
the formation of kerogen is chemical in 
nature and can be further elucidated by 
laboratory tests. 

Components of Ancient Kerogen 

The problem of determining the con- 
stituent groups of kerogen is complicated 
by the insolubility of the substance. Faced 
with this problem, coal chemists have 
adopted a number of procedures, the most 
effective of which is hydrogenation, which 
results in liquefaction of the coal. This 
method is also effective when employed on 
the kerogen of shale as described in last 
year's report and by Forsman and Hunt 

In ordinary hydrogenation of shale, tem- 
peratures of about 380° C are required, so 
that the process is doubtless accompanied 
by some pyrolysis. A method of liquefying 
kerogen at a lower temperature is much to 
be desired. Another goal is to achieve re- 
duction of components of the kerogen to 
hydrocarbons. The present-day gas-liquid 
chromatography devices are extremely ef- 
fective analytical tools in the separation of 
hydrocarbons, and provide a quick and 
sensitive means of detecting and measuring 
various gaseous components. It is reason- 
able to believe that under appropriate con- 
ditions a partly oxygenated carbon chain 
could be reduced to a hydrocarbon of the 
same length. Living matter produces only 
a limited number of types of compounds, 
and the isolation, for example, of «-dodec- 
ane from kerogen after reduction would 
be fairly good evidence that the corre- 
sponding fatty acid had originally been 

These objectives have been accomplished 
in part by use of anhydrous hydrogen 
iodide — a versatile, powerful, reducing 
agent which, moreover, causes breakage of 
ether linkages. Its usefulness in many or- 
ganic chemical reactions was recognized 
nearly a century ago. Raudsepp (1954) 
employed aqueous solutions of the acid in 
studies of Estonian shale and achieved a 
partial hydrogenation. By using anhydrous 
acid and somewhat higher temperatures 
(250° C) a more complete hydrogenation 
and cleavage of kerogen has been obtained 
at the Geophysical Laboratory. The objects 
for study were organic matter from the 
Green River shale and some model systems 
which included a fatty acid, octanoic acid, 
and an amino acid, leucine. 

Crude kerogen was prepared from the 
sedimentary rock by removal of most of 
the inorganic matter. Carbonates were re- 
moved by treatment of finely ground rock 
with HC1 followed by filtration. Siliceous 
matter was decomposed by concentrated 
HF. The residue was treated with HC1, 
filtered, and washed. This procedure im- 
proved the concentration of organic matter 


from 25 to 90 per cent. The kerogen was 
not analyzed, but studies of the Green 
River shale by others have given values of 
C, 61; H, 7.3; O, 28.8; N, 1.4; and S, 1.5 per 
cent. To convert such material to hydro- 
carbon (~C«H 2 n) should require about 
9 g of HI per gram of kerogen. 

Several experiments were performed in 
which kerogen was incubated with excess 
HI at temperatures of 150°, 200°, and 
250° C, for 60, 40, and 8 hours, respectively. 
At 150° C there was much reaction and 
5 g of HI were consumed per gram of kero- 
gen. The kerogen, however, was not ap- 
preciably broken up or rendered soluble. 
At 200° C, 8 g of HI were consumed per 
gram of kerogen and small hydrocarbons 
were produced in 30 per cent yield. Only 
a part of the remaining material was con- 
verted into soluble asphaltenes. At 250° C 
almost all kerogen was converted into 
asphaltenes or into smaller molecules. In 
table 20 is shown a resume of the products 
of the experiment. 

To gain some experience in the behavior 
of the hydrogen iodide reagent, tests were 
carried out employing simple compounds 
as model systems. Heptane and benzene 
were incubated with anhydrous hydrogen 
iodide at temperatures from 180° to 300° C 
for periods of 1 hour to a week in sealed 
glass vessels. Ten milliliters of heptane was 
incubated at 290° C with 44 g HI for 4 
hours. At the end of this time the heptane 
was recovered virtually unchanged. Gas- 
liquid chromatography of the hydrocarbon 
proved that neither isomerization nor 
breakage had occurred in detectable 
amounts. Another run at the same tem- 
perature and duration employing 10 ml 
heptane, 31 g I 2 , and 20 g HI yielded a 
strikingly different result. There was large- 
scale breakage of the heptane, and virtually 
all the hydrocarbons recovered were 
smaller than pentane. Other experiments 
with various mixtures of HI and I 2 indi- 
cated that incubation of carbon chains in 
the presence of I 2 tends to lead to rupture 
of the chain. The presence of a large excess 

of HI tends to diminish the effect of a 
given amount of I 2 . 

In another experiment benzene was in- 
cubated with HI and HI + I 2 . From an 
initial mixture of 8 g benzene and 64 g HI 
incubated for 10 hours at 253° C were re- 
covered 40 per cent benzene, 40 per cent 
cyclohexane, 10 per cent light straight- 
chain hydrocarbons, and 3 per cent tar. In- 
cubation of 8 g benzene with 27 g I 2 and 
75 g HI resulted in formation of ethane, 
propane, and butane. These results indi- 
cate that I 2 is active in splitting C-C bonds 
both in straight-chain and in aromatic link- 

TABLE 20. Hydrogenation of Green River 

Anhydrous HI + kerogen 

250° C, 10 hr, 

Original sample 


Low-temperature volatiles 


High-temperature volatiles 




Organic residue 


Volatiles identified as methane, ethane, pro- 
pane, isobutane (15 per cent), butane (15 per 
cent), isopentane (30 per cent), «-pentane (15 
per cent). 

In the experiments with Green River 
kerogen I 2 is formed as a product of the 
reduction process. Thus, the appearance of 
short straight-chain hydrocarbons among 
the reaction products may well be at least 
in part a result of some scission of C-C 
bonds that were present in the original 

In summary, our results with reduction 
using HI show that this is a means for 
making kerogen amenable to study through 
the production of hydrocarbons and soluble 
asphaltenes. As a by-product of the work, 
the ability of I 2 to break C-C bonds has 
been demonstrated. Detailed interpretation 
of experiments using HI with kerogen 
would require more knowledge of the 
mechanisms of I 2 attack. 



S. P. Clar\, Jr. 

The parts of high-pressure apparatus that 
are subjected to tensile stress are almost 
always made of steel because no superior 
material has been found. In recent years 
new steels that combine high ductility with 
high strength have been developed, and 
such alloys should be well suited to the spe- 
cial requirements of high-pressure equip- 
ment. Tests designed to compare them 
with conventional alloys have been carried 
out; the results indicate that some of the 
new steels are indeed superior to those that 
have been widely used in the past. 

One specimen of each of eleven alloys 
was tested. The steels were selected either 
because they were developed relatively re- 
cently and seemed likely to have desirable 
properties or because they have been used 
previously in high-pressure equipment. A 
single specimen of high-speed steel was in- 
cluded. Hardnesses of the specimens were 
chosen from experience or from manufac- 
turers' recommendations. 

The steels were tested by forcing a ta- 
pered mandrel, made of tungsten carbide, 
into a tapered hole in a cylindrical piece of 
steel. The half -angle of the taper was 3°, 
and the average diameter of the hole was 
0.5 inch. The test pieces were 2 inches in 
diameter and 1% inches long. All tests 
were carried to a permanent enlargement 
of the bore of at least 0.025 inch, or 5 per 
cent of the original mean diameter. They 
were usually terminated when further ad- 
vance of the mandrel led to no further in- 
crease in strength. 

Friction between the mandrel and the 
test piece was reduced by lubricating the 
tapered surfaces with 0.001-inch lead foil 
and with molykote. Nevertheless, more 
than 25 per cent of the total load on the 
mandrel was consumed in friction. This 
could be established from the displacement 
vs. load curve (fig. 58). The mandrel re- 
mained practically stationary as the load 
was reduced from its maximum value (L m 
of fig. 58) to a value that was usually less 

than 50 per cent as great (L r of fig. 58). 
The difference between L m and L r was 
taken to be twice the friction. Hence the 
maximum load actually supported by the 
test piece is (L m + L r )/2; the pressure cor- 
responding to this load can be calculated 
from the dimensions of the tapered hole, 
and these are the values given in table 21. 
The alloys can be divided into four 
groups on the basis of these tests. Strengths 
of about 27,000 bars were shown by Tricent 
and La Belle HT; these alloys were out- 
standing. UHS 260 and Nu Die V were 
nearly as strong, and most of the remain- 

Displocement of mandrel 

Fig. 58. Typical shape of a displacement-load 
curve showing loads L m and L r . 

ing alloys supported pressures in the neigh- 
borhood of 20,000 bars. The high-speed 
steel ruptured at a relatively low pressure, 
a result that was not unexpected; this type 
of steel does not have the ductility required 
for this application, even at relatively low 

Although high strength is generally as- 
sociated with high hardness, no one-to-one 
correspondence between these quantities 
exists. Yield points in simple tension can 
be reasonably well predicted from hardness 
readings, but properties acquired after con- 
siderable cold working cannot be predicted 
from this measurement alone. No values 
of yield points could be obtained from the 
present measurements because the onset of 
plastic flow takes place very gradually in 
this type of test. 



The high strengths shown by the steels 
in these tests are almost certainly not at- 
tainable in routine operations. Experience 

TABLE 21. Alloys Tested and the Results 













La Belle HT 




UHS 260 




Nu Die V 
















Special High- 




13,000 * 

AISI 4140 




Seminole Hard 





AISI 4340 




* Ruptured. 

indicates that any of these alloys will fail 
after a few applications of pressures of 
20,000 bars or more. The relative pressures 

shown in table 21 are of greater signifi- 
cance than the absolute values, which can 
be expected to depend on the type of test to 
a much greater degree. 

The behavior of La Belle HT in service 
has been less satisfactory than these tests 
would indicate. Some of the trouble has 
apparently stemmed from unsound forg- 
ings, but other failures seem to be a conse- 
quence of repeated applications of stress. 
This steel should not be autofrettaged, and 
probably neither should Tricent. Both these 
steels are relatively ductile when their hard- 
ness is considered, but they will not toler- 
ate as much elongation as conventional 
steels that are 10 points softer. 

Nu Die V has been very satisfactory in 
applications where the highest attainable 
strength is not required. Since it is an air- 
hardening steel it undergoes little distor- 
tion on hardening, and intricate shapes can 
be hardened safely. It is also highly re- 
sistant to corrosion, a quality that is de- 
sirable in any steel that must be subjected 
to Washington humidity. 


5. P. Clark, Jr. 

It has been recognized for about sixty 
years that inferences about electrical con- 
ductivity in the earth can be drawn from 
an analysis of transient components of the 
magnetic field. The importance of such 
data has been appreciated only within the 
last ten years or so, however, since little 
was previously known about the electrical 
conductivity of silicates. It is now known 
that the conductivity depends on tempera- 
ture through a Boltzmann factor, and the 
possibility of developing a geothermometer 
based on the electrical conductivity is 

More recently it has been realized that 
heat transfer by radiation may profoundly 
influence the thermal regime of the earth. 
The importance of the process is largely 
determined by the optical absorption co- 
efficient. This raises further interest in the 

electrical conductivity, since certain types 
of electrical conductors absorb strongly. 

Present information about electrical con- 
ductivity in the earth must be regarded as 
preliminary. Large uncertainties exist, and 
there is no way of obtaining an unbiased 
estimate of their magnitudes. It is reason- 
able to suppose, however, that modern ob- 
servations and methods of data processing 
will lead to great improvement in this 
situation, and that a reliable estimate of 
electrical conductivity can be obtained. But 
if much useful information is to emerge 
from the data, the conduction process must 
be thoroughly understood. The extreme 
pressure in the earth inevitably affects the 
electrical properties, and the need to evalu- 
ate the effect provides further incentive to 
gain a fundamental understanding of the 
mechanisms involved. 


Two important conduction mechanisms 
have been recognized in ferromagnesian 
silicates at high temperatures. Above about 
1100° C ionic conductivity overwhelms all 
other types. At lower temperatures a 
mechanism that is believed to be electronic 
in origin is dominant. The relative im- 
portance of the two is unknown at high 
pressure, however. Pressure is known to 
inhibit ionic conductivity, but its effect on 
the electronic mechanism has not been di- 
rectly measured. 

In the following discussion an interpre- 
tation of the electronic structure of silicates, 
based on optical absorption data, is made. 

energy peak is analogous to one found in 
solutions and simple salts containing fer- 
rous iron. It is due to an electronic transi- 
tion within the d levels of Fe ++ . 

The onset of strong absorption at about 
4 ev has been termed the ultraviolet absorp- 
tion edge by analogy with spectra of ele- 
mental semiconductors or simple ionic 
crystals. New data (fig. 59) show that the 
terminology is misleading and that this 
absorption does not arise from electronic 
excitation analogous to intrinsic semicon- 
duction. The absolute values of the trans- 
mission shown in figure 59 have no signifi- 
cance, since all the crystals contained minor 

o o o o 

° ° ° o o O o 


o Synthetic Forsterite 
x Natural Olivine 

4.0 45 

Photon energy, e. v. 

Fig. 59. Optical transmission of forsterite and common olivine in the ultraviolet. 

Considerations pertaining to ionic conduc- 
tion will be left for possible future discus- 
sion. It is advantageous to base an inter- 
pretation on optical spectra because they 
can be understood in terms of crystal field 
theory and because they have been more 
thoroughly studied than the electrical prop- 
erties. But inferences about the conduc- 
tivity drawn from the optical data must be 
regarded as tentative pending experimen- 
tal verification. 

The spectra of ferromagnesian silicates 
have two outstanding features. One is a 
relatively weak absorption peak at a photon 
energy slightly greater than 1 ev, and the 
other is a region of strong absorption at 
energies greater than 4 to 4.5 ev. The low- 

flaws. But the region of almost complete 
opacity which is so obvious in the iron- 
bearing specimen is wholly absent from the 
spectrum of the pure magnesian crystals. 
The presence of iron is essential to this 

The most plausible explanation of this 
observation is that the excitation of an elec- 
tron from an oxygen ion to the lower (t 2g ) 
level occupied by d electrons in a Fe ++ ion 
causes the absorption. Such transitions have 
been found in many inorganic complexes; 
they cause what is sometimes referred to as 
charge-transfer spectra. The maximum ab- 
sorption coefficient is expected to be 
roughly 1000 cm" 1 , which is sufficient to 
cause the observed opacity of crystals 



Y 4 mm or more thick. This interpretation 
is also consistent with the observed fact 
that this region of absorption shifts to- 
wards the red under pressure. 

It is tempting to associate the charge- 
transfer spectrum with the electronic con- 
ductivity. The thermal excitation energy, 
determined from measurements of conduc- 
tivity as a function of temperature, is about 
3 ev compared with the optically deter- 
mined energy of about 4.5 ev. The dis- 
crepancy need cause no concern, however, 
since optical energy gaps are always greater 
than thermal gaps. An attractive feature of 
this interpretation of the conductivity is 
that it provides a natural explanation of the 
low carrier mobility which has been in- 
ferred from conductivity measurements. 
The excitation of an electron to a d level 

of iron will usually lead to the formation 
of bound pairs of electrons and holes, 
which cannot contribute to conductivity. 
Only rarely will these be able to separate, 
and as a consequence the mobility will, on 
the average, be very low. 

Conclusions drawn from this model of 
the electronic structure of ferromagnesian 
silicates must be regarded as tentative, but 
the identification of energy levels has con- 
siderable value as a guide to future experi- 
mentation. The development of a tech- 
nique for the preparation of much thinner 
crystals than have previously been studied 
will greatly enhance the value of optical 
studies. But more important will be elec- 
trical and optical studies on synthetic crys- 
tals of variable and controlled iron content. 


G. Donnay and J. D. H. Donnay 10 

The first report given on this subject 
(Year Book 56, pp. 244-246) told of a 
largely unexplored field — the application of 
generalized symmetry theory to the deter- 
mination of magnetic crystal structures — 
which invited further study. In addition to 
clarifying a dubious point in the derivation 
of the Shubnikov groups (Donnay et al., 
1958), we investigated the following prob- 

Tables of magnetic space groups. To 
facilitate the use of symmetry criteria as 
an aid in magnetic structure determination, 
the tabulation of symmetry properties that 
already exists for the 230 Fedorov-Schoen- 
flies groups should be extended to the 1421 
generalized space groups that are possible 
for ferro-, ferri-, and antiferromagnetic 
structures. At first sight the increase that 
would seem to be required in the size of 
the tables appears formidable, but a con- 
densed scheme of presentation is in prepa- 
ration (Donnay and Donnay, 1959), which 
will give all the essential data and yet will 
not require more than a small brochure, to 

The Johns Hopkins University. 

be used in conjunction with the present 
International Tables. 

With the help of the ten theorems of 
Belov et al. (1955), it is possible to con- 
struct a figure for any group with antisym- 
metry elements, from the diagram of the 
corresponding Fedorov-Schoenflies space 
group to be found in the International 
Tables. The generalized point-group sym- 
metry of any site can then be derived by 
inspection. If, and only if, it is a subgroup 
of oo fm 2' fm' (the generalized symmetry 
of an axial vector) can the site be occupied 
by a magnetic atom. (The magnetic mo- 
ment is here considered a dipole, hence an 
axial vector.) Once the axial vector has 
been chosen in direction, sense, and mag- 
nitude, for one of the sites of a position, the 
axial vectors on all the other sites are fixed 
by symmetry. If the position is a special 
one, so that the point-group symmetry of 
the site is different from 1, the axial vector 
will have to lie along the symmetry direc- 
tion (direction of an axis of symmetry or 
perpendicular to a plane of symmetry) so 
that, for a given magnitude, only its sense 
can be selected at will on one of the sites of 
the position. Thus, for special positions, 
the axial vector can be located in space on 



inspection. A tabulation is needed, how- 
ever, for the general position; each site has 
symmetry 1, and the direction [uvw] of 
the axial vector is general, so that the deri- 
vation of the related axial vectors on the 
remaining sites is tedious. It can be given 
in table form for all the space groups that 
are derived from a given symmetry point 
group and its related antisymmetry point 

include any glide component, the n sites 
are distributed around the origin (this is 
the case in Fedorov's symmorphic groups). 
If, on the other hand, the n sites are scat- 
tered in the cell by glide reflections or 
screw rotations, it is possible, by applying 
the glide components in the opposite sense, 
to bring them back around the origin and 
to retrieve the distribution of the sym- 

TABLE 22. Distribution of Axial Vectors on the Sites of the General Position of All Space 
Groups Derived from Point Group 4/m 

Fractional coordinates 


sites in general 



position, referred to 

origin at 


, and 






glide components to 







sponding coordinates 


the various 

















Symmetry of As- 

semblage of 



Moments * on 





: 000 

000; HI 

Polar Axial 



d A 


C D 


Vectors Vectors 

4(x2) 4/m 




4 1 


3 4 





IA/m lAJa 


4(x2) A'/m 




IV 1 


3 IV 

PA' /m 

PA 2 '/m 

PA' In 

PA 2 '/n 

lA'/m IA ± '/a 


A/m A/m' 




4 I 







IA/m' lAJa' 


A/m=A/m A'/m' 






III 4 

PA' lm' 

PA 2 '/m' 

PA' /n' 

PA 2 '/n' 

IA'/m' lA^/a' 



000; HI; 


000; 00JR 

001R; IIQR 

P c A/m 

P c %/m 

P c A/n 

P c 4Jn 

/ c 4/m J„V fl 


4(x2)(x2) A/ml' 




4 1 


3 4 


000; 1I0R 

P c A/m 

P AJm 

P c A/n 

P c A 2 /n 



000; IUR 


PjA 2 /m 




Legend of moments: 1, {uvw\; 2, [vuw\; 3, [tivw]; 4, [vuw]; I, [«»«/]; II, [vuw]; 

']; IV, [ftiW]. 

groups. A sample for point group 4/m is 
shown in table 22. The required 32 tables 
are now being prepared jointly with Dr. 
N. V. Belov and his co-workers. 

Alternative derivation of the magnetic 
space groups. Consider a space group. Let 
m be the multiplicity of the general posi- 
tion. It is known that m — tn, where t is 
characteristic of the centering or anticenter- 
ing of the cell and n is the number of 
equivalent sites that result from symmetry 
operations other than lattice translations or 
antitranslations. If these operations do not 

morphic groups. We shall call it the collec- 
tion of sites around the origin. Its point- 
group symmetry can be expressed by one 
of the 32 Hessel-Bravais groups. 

Let us now consider the magnetic mo- 
ments, taken to be dipoles or axial vectors, 
associated with the n sites. Let them pass 
through one and the same point. The as- 
semblage can be represented on the sphere 
of projection drawn around the point of in- 
tersection of the vectors; its point-group 
symmetry will be one of the 122 Heesch 
groups. Now if only the sense of the axial 



vectors is taken into account, that is to say, 
if these axial vectors are supposed to be 
transformed into polar vectors, the point- 
group symmetry of the new assemblage 
can be expressed as one of the 32 groups of 

The following rules (see table 23) enable 
one to pass from the Heesch group of the 
collection of axial vectors to the Hessel- 
Bravais group of the collection of polar 
vectors: all symmetry axes, inversion axes 
n as well as rotation axes n, become rota- 
tion axes n; all antisymmetry axes, n 
or ri ', become inversion axes n. (These 

TABLE 23. Correspondence Between Symmetry 
Elements in the Two Assemblages of Vectors 

Assemblage of Polar 




Assemblage of Axial 


(Heesch group) 


2=1/ m 

1 or _ 1 

2 or 2=l/m 
n or n 

V or _ T' 

2' or 2'=l/m' 
ri or n 

Legend: 1, identity; 2, 2-axis; n, rotation axis; 
1, center; m, mirror; n, inversion axis; 1', anti- 
identity; 2', 2'-antiaxis; ri, rotation antiaxis; m', 
antimirror; ri, inversion andaxis. 

rules follow from the following facts: (1) 
an axial vector remains invariant under in- 
version of space; (2) the time reversal R, 
which is symbolically represented by the 
change from black to white in the "black- 
white" groups, reverses the sense of the 
current in the loop of the axial vector, and 
this corresponds to the inversion of the 
"polar vector.") 

The derivation proceeds in two steps. 

1. For every Laue class all the point- 
group symmetries of the collection of sites 
around the origin are combined with the 
same point-group symmetries of the collec- 
tion of polar vectors (arrows) . The former 
are used for row headings ("sites"), the 
latter for column headings ("arrows"). As 
an example let us take Laue class 4/m 

(table 24). To obtain all the possible 
Heesch groups for the collection of axial 
vectors, we first write the point-group sym- 
metry of the collection of sites around the 
origin (to which we may add identity 1) 
in every pigeonhole of the corresponding 
row of the table. Then, by adding primes 
(according to the rules shown in table 23), 
we substitute antisymmetry elements for 
symmetry elements of the second kind — 
inversion axes, including mirrors and cen- 
ters — that appear in the column headings. 
In the column headed 4/m we must change 
identity 1 to anti-identity 1', which is the 
only way to double the order of the group; 
the result is a "gray" Heesch group, in 

TABLE 24. Derivation of Heesch Groups in 
Laue Class 4/m 















which each site carries two axial vectors 
with opposite senses. 

If both assemblages ("sites" and "ar- 
rows") are centrosymmetric, more than 
one Heesch group is obtained. As shown 
in table 23 the antielements can be intro- 
duced in more than one way. One may 
add 1' and obtain a gray group as for the 
noncentrosymmetric symmetries of the col- 
lection of sites; one may replace the mirror 
by an antimirror; finally, by considering 
that the symmetry 4/m of the assemblage 
of arrows can also be written 4/m, one ob- 
tains the Heesch group 4'/m'. 

The 122 Heesch groups can thus be de- 
rived from the 32 classical point groups in 
11 tables, one for each Laue group. We 
propose adding these 11 tables to the bro- 
chure that will contain the tables of mag- 
netic space groups. 

2. For each symmetry of the collection 



of sites in turn, we can now derive all the 
space groups related to it. The same 32 
tables proposed for the brochure can be 
used here. Let us consider group 4/ra 
(table 22). Two kinds of space groups 
will be derived, according as the lattice is 
one of the Bravais lattices (here, P or /) or 
one of the antilattices (here, P c , Pc, Pi, or 
Ic). Note that we assume the generalized 
lattices as a prerequisite to derive the gen- 
eralized space groups, just as the Bravais 
lattices are a prerequisite to the derivation 
of the 230 classical space groups. It follows 
from Belov's theorems (1955) that the non- 
gray Heesch groups for the assemblage of 
axial vectors can be combined only with 
Bravais lattices, whereas the gray groups 
can be combined only with antilattices; 
they give nongray generalized space 
groups. The first distribution of axial vec- 
tors on the sites can occur in all the classi- 
cal space groups isomorphic with 4/m 
(table 22, row 1). The second distribution 
corresponds to an assemblage of axial vec- 
tors with Heesch group 4'/ra; the corre- 
sponding space groups have the same sym- 
bols as those in row 1, except that 4 is 
replaced by 4' (table 22, row 2). Likewise 
we obtain the space groups in rows 3 and 4 
(table 22) by appropriate priming of the 
antielements. All space groups isomorphic 
with 4/ra have now been derived. 

Next the antilattices P c are combined 
with the gray Heesch group 4/ral' (table 
22, row 5). Of the two axial vectors with 
opposite senses to be found on each site in 
the collection of sites around the origin, 
one remains in place, the other is carried 
to the corresponding site around 00|. 
Thus the general position in a space group 
with an antilattice has twice the multi- 
plicity of the corresponding one in a space 
group with Bravais lattice, and any space 
group with antilattice is derived from a 
gray point group. The remaining space 
groups with antilattices h, Pc, and Pi are 
obtained in analogous manner (table 22, 
rows 5, 6, and 7) . 

The distribution of spins in magnetic 
crystal structures of the sodium chloride 

type. The interest shown by physicists in 
the spin arrangement of the antiferromag- 
netic state of "cubic" crystals of the NaCl 
structural type prompts us to offer, as a 
crystallographic contribution to this prob- 
lem, an analysis of the requirements of 
symmetry and antisymmetry acting on an 
atom that carries an axial vector. The 
seven ferri- and antiferromagnetic com- 
pounds so far investigated are MnO, FeO, 
CoO, NiO, MnS, MnSe, CrN. Cubic above 
the Neel point, their symmetry is lowered 
and their cell edge doubled below it. Loeb 
(1958), without taking symmetry into ac- 
count, derived relations, for permitted spin 
directions, which are based on the require- 
ment that the magnetic dipole interactions 
should have minimum energy. He found 
the solution to be nonunique, since he ob- 
tained eight equations with twelve varia- 
bles. These equations can be used, how- 
ever, to test a given hypothesis. It has long 
been known that, on the assumption of ex- 
change forces, the doubling of the cubic 
cell edge can be predicted. 

If the symmetry of the magnetic crystal 
structure is assumed to remain rigorously 
cubic, the symmetry analysis leads to three 
distinct spin distributions provided the cell 
edge of the magnetic cell remains equal to 
that of the chemical cell; if the cell edge is 
doubled, only one spin distribution is pos- 
sible. The spin directions belong to <111> 
in all cases. None of these cubic spin dis- 
tributions satisfies Loeb's equations, not 
even the spin distribution for the larger 
cell. Cubic magnetic structures of the NaCl 
type may be considered unlikely on the 
basis of energy considerations; from the 
point of view of symmetry alone, however, 
we have shown that they are not ruled out. 

N. Morimoto 
The pyroxenes in the field CaMgSi 2 Oo- 
CaFeSi 2 0«-MgSi03-FeSi03, which are the 
most important pyroxenes in nature, can 
be divided into two different types. One 
group is near the CaMgSi 2 06-CaFeSi20c 
join, with more than 25 mole per cent 


CaSi0 3 component (a type) ; and the other 
group lies near the MgSi0 3 -FeSi03 join 
with less than 15 mole per cent CaSiOa 
component (b type) . 

The fl-type pyroxenes exist only in a 
monoclinic form and belong to the space 
group C2i/c (C271 6 ). Diopside is considered 
to have the type structure of this group. 

The b-type pyroxenes have been known 
in two forms in nature, monoclinic and 
orthorhombic. The pyroxenes of this type 
with more than about 30 per cent FeSiOs 
have the monoclinic form, pigeonite, or 
the orthorhombic form, hypersthene. The 
mode of occurrence clearly shows that 
pigeonite is a high-temperature form and 
hypersthene a low-temperature form (Pol- 
dervaart and Hess, 1951). When the 
amount of FeSi0 3 component is less than 
about 30 per cent the thermal experiments 
confirm the existence of three polymorphs 
— enstatite, protoenstatite, and clinoensta- 
tite (Foster, 1951) — although only the or- 
thorhombic form is known in nature except 
in some meteorites. The problem of the sta- 
bility relations among these three poly- 
morphs has not yet been settled. Foster 
(1951) and Atlas (1952) have proposed the 
following relationship : protoenstatite, 
high-temperature form; enstatite, low-tem- 
perature form; and clinoenstatite, low-tem- 
perature metastable form. The relationship 
between pigeonite and hypersthene, how- 
ever, suggests that possibly clinoenstatite 
is a metastable high-temperature form of 
enstatite. The monoclinic forms of the 
b-type pyroxenes, clinoenstatite and pigeon- 
ite, have the space group P2i/c (C211 5 ) and 
are distinct from the a-type pyroxenes, 
which have the space group C2i/c (Mori- 
moto, 1956; Bown and Gay, 1957). 

On the basis of these pyroxene relations, 
our efforts during this past year have been 
concentrated on the following two crys- 
tallography problems: the structural rela- 
tions between diopside, pigeonite, and 
clinoenstatite; and the structural relations 
between the three polymorphs of MgSiO- 
— enstatite, protoenstatite, and clinoensta- 
tite. The structural relation between en- 

statite and ferrohypersthene is also being 
studied to clarify the effect of the replace- 
ment of Mg by Fe on the structure of the 
orthorhombic pyroxenes. 

The Structural Relations between Diop- 
side, Clinoenstatite, and Pigeonite 

N. Morimoto, D. E. Appleman, 6 and 
H. T. Evans, Jr. 6 

The crystal structures of clinoenstatite 
and pigeonite have been determined to 
elucidate the structural difference between 
the clinopyroxenes of the a and the b types 
and to understand the effect of substitution 
among Fe, Mg, and Ca atoms on the single 
chain silicate structures. 

The specimens of clinoenstatite were ob- 
tained by heating Bishopville enstatite at 
1400° C for 24 hours. The cell dimensions 
are a = 9.618 ±0.005, b = 8.825 ±0.005, c- 
5.186±0.005 A, and (3 = 108° 21'±5'. The 
specimens of pigeonite were kindly do- 
nated by Professor H. Kuno, who collected 
them from andesite and dacite dikes near 
Asio Mine, Japan, and stated the chemical 
composition to be Woi2En 2 6Fs62 (Kuno, 
personal communication, 1959). The cell 
dimensions are a = 9.731 ± 0.005, b = 8.953 ± 
0.005, c = 5.256 ±0.005 A, and 3=108° 33' 

Starting with the structure of diopside, 
published by Warren and Bragg (1928), 
the Fourier syntheses, the (Fo-Fc) syn- 
theses, and the least-squares method have 
been applied using (A^0), (hOl), and 
(0^/) data for both crystals. The final re- 
liability factors for these reflections are 
about 10 per cent in both structures. 

The results reveal details of these struc- 
tures which are based on the single silicate 
chain characteristic of the pyroxene group. 
The projections on the (010) plane of the 
structures of clinoenstatite and pigeonite 
finally obtained are given in figure 60 and 
compared with the projection of diopside. 
The results of interest obtained in this 
study are described below. 

Metal atoms. There are two crystallo- 
graphically different positions for metal 



atoms, Mi and Mn, in the monoclinic 
pyroxene structure. Table 25 gives the 
distribution of metal atoms in the Mi and 
Mn positions, the coordination numbers 
of the metal atoms, and the average atomic 
distances from the metal atoms to the 
neighboring oxygen atoms. It is clear from 
the table that metal atoms Ca, Fe, and 
Mg are in completely ordered distribution 
in the Mi and Mn positions, with the 
larger atoms in the Mi position by prefer- 
ence. If enough larger atoms are not avail- 
able, small atoms occupy this position, and 
the coordination number tends to change 
from 8 to 6. The coordination of the Mi 
and Mn positions is shown in figure 60 
with dotted lines for Mn and dotted and 
dashed lines for Mi. 

are built up from the single silicate chains 
found in diopside, there are some differ- 
ences in the details of the chain structures. 
The silicate chains in the three clino- 
pyroxenes are projected perpendicular to 
(100) in figure 61, where the plane passing 
through the Si atoms is taken as the zero 
level and the heights of oxygen atoms from 
the zero level are represented by Ang- 
strom units. There is only one kind of 
chain in diopside, (E) in figure 61, because 
all chains must be crystallographically 
equivalent. There are, however, two crys- 
tallographically different kinds of chains in 
clinoenstatite and pigeonite. In figure 61, 
(A) and (B) are chains of clinoenstatite 
and (C) and (D) are those of pigeonite. It 
is interesting to note the constant distances 

TABLE 25. Nature of the Metal Atoms in Clinoenstatite, Pigeonite, and Diopside 

Coordi- M— O 
Atoms nation Atomic 

Number Distance 



Diopside (Warren and Bragg, 1928) 









M x 

^ a 0.24^ e 0.76 




Mgo.52Fe . 48 



M r 








Both Mi and Mn positions are on the 
twofold axes in the diopside structure, but 
in clinoenstatite and pigeonite the twofold 
axes are absent and they lie in general posi- 
tions. Metal atoms at the Mi and Mn posi- 
tions in diopside, therefore, exactly super- 
pose in the projection on (010), but in 
pigeonite and clinoenstatite they show 
some separation in that projection. The 
deviations of the coordinates of the Mi and 
Mn positions in clinoenstatite and pigeon- 
ite from those of diopside are given in 
table 26 in Angstrom units. The deviation 
of Mi from its location in diopside is 
greater than that of Mn, and the deviations 
of Mi and Mn are greater in clinoenstatite 
than in pigeonite. 

The single silicate chains. Although the 
structures of clinoenstatite and pigeonite 

(3.04 ±0.02 A) between neighboring Si 
atoms belonging to the same chain in all 
three structures regardless of the kind of 
metal atoms. The difference of the shape 
of the chains, therefore, is caused by some 
rotation of Si0 4 tetrahedra around the axis 
passing through Si and OI atoms without 
changing the relative position of Si atoms. 
The direction of rotation is opposite for the 
SiC>4 tetrahedra belonging to two crystallo- 
graphically different chains in the same 

The main effect of the metal atoms on 
the single chains, however, can be observed 
in the displacement of silicate chains along 
the c axis rather than in the change of 
atomic positions in the chain itself, in order 
to make an appropriate coordination of 
oxygen atoms for the metal atoms as shown 


(a) Clino-enstatite 

(b) Pigeonite 

(c) Diopside 

Fig. 60. The structure of clinoenstatite (a), pigeonite (b), and diopside (c) projected on (010). 
The numbers give the height of each atom in the cell, expressed by a percentage of the ^-translation. 
The structure of diopside is taken from Warren and Bragg (1928). The coordinations of the M n 
atoms are represented by dotted lines and those of M l by dotted and dashed lines. 


TABLE 26. The Displacement of Metal Atoms 


M T 

MgSi0 3 


(Mg 2 6Fe 62 Ca 12 )Si0 3 
(Ca 50 Mg 50 )SiO 3 

Mg 100 

re 76 Ca 24 

^ a ioo 
(0, 0, 0) 


Mg 52 Fe 48 

Mg 100 

* The values of deviation from the position of Ca in the diopside structure are given by Angstrom 
units in the order of the direction of the a, b, and c axes. 

in figure 60. The changes both of the shape 
and the relative position of the chains are 
greater in clinoenstatite than in pigeonite 
compared with the diopside chains. 

The relation between clinoenstatite and 
pigeonite. In structure, pigeonite is inter- 
mediate between clinoenstatite and diop- 
side. The seven coordinations of the atoms 
in the Mi position, instead of six in clino- 
enstatite or eight in diopside, and the devia- 
tion of the Mi and Mn positions, which is 
just between diopside and clinoenstatite, 
are especially characteristic. 

We do not know, however, whether all 
the atoms in Mi positions in pigeonite 
really have seven coordinations or the 
seven coordinations are only a statistical re- 
sult of the eight coordinations around Ca 
atoms and the six coordinations of Fe 
atoms at the Mi position. Neither do we 

know whether all the Mi and Mn positions 
are displaced from those of diopside or the 
deviation of Mi and Mn positions is a sta- 
tistical result of deviation of Fe atoms and 
nondeviation of Ca atoms. 

It is difficult to decide which interpreta- 
tion is better at the present stage, especially 
because of poor resolution of the electron 
density in the (010) projections of the three 
structures. The diffuseness of the addi- 
tional reflections (h + l^ = 2n + l) in some 
pigeonites (Bown and Gay, 1957) could be 
explained on the basis of the local adjust- 
ments of coordination about the metal 
atoms as suggested in the last paragraph, 
supplemented by the incipient exsolution 
to the Ca-rich regions and the Fe,Mg-rich 
regions (Morimoto and Ito, 1958). 

The structure relationship of clinopy- 
roxenes. Although the crystal structure of 

(A) (B) 



Fig. 61. The single silicate chains in clinoenstatite (A) and (B), pigeonite (C) and (D), and 
diopside (E), projected perpendicular to (100). The planes passing through the Si atoms are taken 
as the zero level, and the heights of oxygen atoms from the zero level are represented by Angstrom 


clinopyroxene is principally built up of 
single silicate chains, slight changes of 
these chains take place depending on the 
kinds of metal atoms in the structures. The 
change of structures of clinopyroxenes can 
conveniently be described by the following 
three cases, taking the diopside structures 
as standard: (1) displacement of the sili- 
cate chains, (2) distortion of the silicate 
chains, and (3) displacement of the metal 

is characteristic in the diopside structure, 
producing two kinds of silicate chains. 
Thus the space group becomes P2i/c. The 
clinopyroxenes with compositions along the 
enstatite-ferrosilite join are examples of 
this case, 3. They are no longer stable in 
nature, except under special conditions, and 
they transform to orthopyroxenes, which 
are stable at low temperature, exsolving 
Ca atoms from the structures as diopsidic 

TABLE 27. Crystallographic Data and Stability Relations for Three Polymorphs of MgSiO s 





217=835.8 A 3 


[3=108° 30' 


a— 18.230 


low-temperature form 


7=832.0 A 3 





high-temperature form 

(Smith, 1958) 


217=860.2 A 3 

The clinopyroxenes of the compositions 
along the diopside-hedenbergite join are 
examples of case 1. Although all clino- 
pyroxenes of this group have practically the 
same translation along the c axis, which is 
the direction of the silicate chains, the dis- 
tance between the silicate chains changes 
to make appropriate space for the metal 
atoms of various sizes. This results in a 
slight change of the cell constants, except 
the c translation. 

When the displacement of the silicate 
chains is not sufficient to give space for 
metal atoms, a distortion of silicate chains 
takes place as in spodumene (Warren and 
Biscoe, 1931). Various augites show the 
effect of case 2. The clinopyroxenes of the 
above two cases have one kind of silicate 
chain and have all metal atoms on the two- 
fold axes, belonging to the space group 
C2i/c. The change in position or distortion 
of the silicate chain is inadequate to com- 
pensate for the introduction of small diva- 
lent cations such as Fe or Mg atoms into the 
Mi positions. The Fe or Mg atoms become 
displaced from the special position, which 

The Structural Relations among Three 

Polymorphs of MgSi0 3 — Enstatite, 

Protoenstatite, and Clinoenstatite 

N. Morimoto 

In order to understand the nature of the 
transition between the three polymorphs 
of MgSiO.3, it is necessary to know their 
structures accurately. The crystallographic 
data for the three polymorphs are given in 
table 27. The structure of protoenstatite 
has been determined by Smith (1958) us- 
ing powder data. The refinement of the 
structure of enstatite has been undertaken 
using Bishopville enstatite single crystals. 

The difference of the atomic positions in 
the (001) projection between enstatite and 
clinoenstatite structures is less than 0.1 A. 
As the projection on (001) of the structure 
of protoenstatite is also very close to the 
projections of enstatite and clinoenstatite, 
all the atoms in the three polymorphs have 
practically the same x and y coordinates. 
The difference in their structures, there- 
fore, depends only on the z parameter of 
atoms. In figure 62, the projections on 
(010) of enstatite and clinoenstatite struc- 



tures are shown and compared with the 
projection of protoenstatite. 

The following facts appear from the 

1. The most significant difference in the 
three structures is the position of Mg 

(a) Clino-enstatite 

(c) Proto-enstatite 

O o • 

Mg Si 

Fig. 62. The structure of clinoenstatite (a), 
enstatite (b), and protoenstatite (c) projected on 
(010). The light lines represent their unit cells. 
The structure of protoenstatite is taken from 
Smith (1958). 

atoms. All Mg atoms in protoenstatite lie 
on planes parallel to (100) at z and z + Y 2 
coordinates. The position of Mg atoms is 
shifted, however, by % of the c translation 
in every other plane parallel to (100) in 

clinoenstatite and in every other pair of 
planes in enstatite. If we represent the plane 
of Mg atoms by A or B according as the 
metal coordinates are z and z + 1 / 2 , or z + % 
and z + %, the structures along the a axis 
can be expressed as follows : protoenstatite, 
AAA • ■ ■ ; enstatite, A ABB A ABB- ■ • ; and 
clinoenstatite, ABABA ■ • • . Their struc- 
tures are connected, therefore, by a glide of 
Y 4 of the c translation parallel to (100), 
which is the twin plane of clinoenstatite. 

2. The shape and the position of the sili- 
cate single chains in the; structures of the 
three polymorphs depend on the relative 
position of Mg atoms, giving an appropri- 
ate octahedral coordination of O atoms 
around Mg atoms. The three polymorphs 
have, therefore, chains slightly different 
from one another. 

3. The transitions between the three 
structures can be reasonably explained by 
the displacement of Mg atoms without the 
breaking of any silicate single chain. The 
transitions themselves, however, should be 
considered reconstructive, inasmuch as 
some of the bonds between Mg and O 
atoms must be broken on transition. 

If mistakes take place in the arrange- 
ment of Mg atoms in the structure of 
MgSiOs, intermediate forms appear. In 
fact, some intermediate forms are found 
in dry experiments in the ternary system 
MgSi0 3 -diopside-albite (see Schairer and 
Morimoto, this report) . Their powder pat- 
terns show them to be different from any 
of the three polymorphs. After a couple of 
weeks, however, the patterns approach the 
pattern of natural enstatite, showing that 
intermediate forms are not necessarily 

CaAl 2 Si 2 8 

G. Donnay and Y. Ta\euchi xx 

In a previous note (Donnay, 1952) on 
synthetic hexagonal CaAl 2 Si 2 8 the crys- 
tal data were reported and attention was 
drawn to the fact that the crystal structure 

11 University of Tokyo. 


is not the same as that of BaAl 2 Si 2 8 (Ito, 
1950), despite a similarity in cell geometry 
and space group which would suggest an 
isostructural character. Dr. Takeuchi, to- 
gether with G. Donnay, has determined 
the crystal structure by two-dimensional 
Fourier methods. A final residual R of 
13.8 per cent takes into account structurally 
absent reflections. The structure consists of 
double sheets of (Si,Al) tetrahedra parallel 
to (0001). Symmetry requires silicon and 
aluminum to replace each other randomly. 
All tetrahedra in the lower half of the layer 
point up, those in the upper half point 
down; the angle of 180° at the apical 
"bridge oxygen" (oxygen shared between 
two tetrahedra) is also required by sym- 
metry. Because all corners of the tetrahedra 
are shared, the resulting (Si + Al)/0 ratio 
equals that of a framework structure, and 
the formula is the same as that of anorthite. 
These results lead to the following con- 

1. The presence of 50 per cent of Al in 
the tetrahedra is obviously sufficient to 
open the bond angle at one of the bridge 
oxygens to the value of 180°, expected only 
for a purely ionic bond. In aluminosilicates 
with less aluminum in tetrahedral coordi- 
nation, this angle is equal to 145° ±15°, a 
range which supports the hypothesis that 
the Si-O bond is approximately half ionic, 
and half covalent in character, with a pre- 
dicted bridge bond angle of (180° + 
109 ° ) /2 = 144 1 / 2 ° . Thus we are led to think 
that the value of the bond angle at a bridge 
oxygen in an aluminosilicate can yield in- 
formation on the extent of (Si,Al) substi- 
tution. Until now only the size of the 
tetrahedra, that is the average (Si,Al)-0 
bond length, had been used to make such 

2. Compared with the structures of the 
low- and high-temperature forms of the 
corresponding barium compound (deter- 
mined by Takeuchi, 1958), the structure of 
the calcium compound shows major differ- 
ences in the configuration of the double 
layers. Only the radius of the cation (Ca ++ , 
Ba ++ iow temp., or Ba ++ hi K h temp.) can be held 

responsible for these differences. Thus the 
alkali-earth ion determines the detailed 
structure. Yet the calcium coordination 
polyhedra are found to be of two kinds. 
If adjacent double layers are turned 60° 
with respect to each other, every calcium 
ion finds itself at the center of an almost 
regular octahedron of oxygen atoms and 
the structure is an ordered one. Diffuse 
reflections on the X-ray diagrams, however, 
indicate that mistakes in the stacking of 
layers are frequent. The coordination poly- 
hedron of a calcium ion between adjacent 
double layers which are not turned with 
respect to each other (or which are turned 
120° or 240°) is a trigonal prism. Thus, in 
predicting new silicate structures, it is not 
safe to assume a single cation coordination, 
which would fix the skeleton of a silicate 
structure, and accordingly to adjust the 
aluminosilicate building blocks to their 
oxygen environment. Such an approach, 
which is being used nowadays, may be 

3. The absence of a high-temperature 
form for the calcium compound is made 
plausible by a comparison of the same 
structures. In the low-temperature form of 
BaALSiaOg the layer of tetrahedra consists 
of six-membered rings with ditrigonal, 
pseudohexagonal symmetry; these rings 
acquire perfect hexagonal symmetry in the 
high-temperature form. In CaAUSi-O? the 
ring is ditrigonal, no longer pseudohex- 
agonal, and the change to hexagonal sym- 
metry would have to be a much more 
drastic one; it has not been observed. 

N. Morimoto and G. Kullerud 

Single crystals of this phase were syn- 
thesized by two methods: (1) by heating 
synthetic FeSe 2 (ferroselite) in evacuated 
silica tubes at 600° C, where it readily de- 
composes to Fe 3 Se4 and vapor; (2) by 
heating FeSe with appropriate amounts of 
Se at the same temperature. Heating for 
7 days by the first method produced crystals 
large enough for single-crystal work, 



whereas 20 days were required by the sec- 
ond method. 

The two methods of synthesis demon- 
strated the existence of measurable solid 
solution of Se in this phase. By the first 
method crystals in equilibrium with free 
Se were produced. These crystals contained 
more Se than is given by the FesSe4 for- 
mula. Comparison of X-ray diffraction 
powder patterns of material synthesized by 
the two methods showed shifts of the 
X-ray reflections due to differences in com- 

The crystals range in size from 0.05 to 
0.5 mm. Most of them are euhedral hex- 
agonal plates which represent a combina- 
tion of a well developed base {0001} and a 
rhombohedron. The crystals usually occur 
in aggregates of intersecting plates. No 
prism faces were observed. Subhedral ir- 
regularly formed crystals are uncommon. 
One crystal of dimensions about 0.2x0.2 
X 0.1 mm 3 was studied by the two-circle 
optical goniometer. On the basis of the 
rhombohedral axis, this crystal has the 
forms {111} and {100}. The point sym- 
metry is 3tn, and a = 81° 38' with (111) A 
(100) =60° 54'. Although the X-ray dif- 
fraction powder patterns show the general 
features of a NiAs structure, additional 
peaks occur that cannot be accounted for 
on the basis of hexagonal symmetry. 

Single-crystal studies were performed on 
the same crystal that was used for morpho- 
logical measurements. Weissenberg and 
precession photographs were taken with 
Mo Ka and Cu Ka radiation. The pre- 
cession photographs taken parallel to the 
basal plane of the crystal, (0001)^*, show 
the hexagonal symmetry (subscript h re- 
fers to the hexagonal axes). In the higher 
layer photographs, however, each reflection 
splits into three, suggesting a multiple 
twinning of three edifices with a pseudo- 
hexagonal symmetry. 

All photographs can be interpreted on 
the basis of multiple twinning of three edi- 
fices having a common basal plane but 
with a 120° difference in orientation. Each 

edifice has a pseudohexagonal, truly mono- 
clinic lattice with 3 = 92° ±0.5°. The three 
edifices have identical volumes, to judge 
from the intensities of the reflections. This 
type of twinning suggests that the crystals 
possessed a hexagonal symmetry at the 
temperature of growth (600° C), consistent 
with the conclusions from the morphologi- 
cal observations. 

Cry stallo graphic data. On the basis of 
the interpretation of the twinning mecha- 
nism, the following crystallographic data 
were obtained for material grown as de- 
scribed under method 1. Since this ma- 
terial contains more Se than FesSe4, its for- 
mula may be given as Fe3-#Se4. a — 6.059 ± 
0.005, £ = 3.491 ±0.002, c = 10.95 ±0.01 A; 

TABLE 28. The Atomic Coordinates of FeSe* 

Atoms * 




























* General position for C 2h 3 — 12/ m: (000, \ 

2 h) -f- xyz, xyz, xyz, xyz. 

3 = 92° ±0.5°; space group C2h 2 — I2/m; 
Z = 2(Fe 3 -a;Se4). The cell dimensions, 
which were refined by the X-ray diffrac- 
tion powder method, are slightly smaller 
than those reported by Okazaki and Hira- 
kawa (1955) for stoichiometric FesSe4. 

Structure. Okazaki and Hirakawa 
(1955) have proposed a structure derived 
from FeSe, which has a distorted NiAs- 
type structure, by subtracting 1 Fe from 

4 FeSe to obtain Fe3Se4. The subtraction 
of Fe atoms results in vacancies which are 

To check this proposed structure, the 
structure factors of the (^00) reflections 
were calculated using the atomic coordi- 
nates given in table 28. For the calculation 
a temperature factor of B—\5 was used. 
The observed and calculated values are 
compared in table 29. The good agreement 
between Fobs, and F ca \. confirms the or- 


TABLE 29. Comparison of Observed F's and 
Calculated F's 


sin Q/X 

F cal .(5=1.5) 




















- 8.9 






dered arrangements of the Fe atoms in the 
distorted NiAs-type structure. 


N. Morimoto and G. Kullerud 

Single crystals of digenite and bornite 
and of various compositions of the digen- 
ite-bornite join were synthesized during a 
current study of the subsolidus relations 
between CU9S5 and Cu 5 FeS4. 

The twinning mechanism and the crys- 
tallographic properties of rhombohedral 
digenite were studied by Donnay, Donnay, 
and Kullerud (1958). High-temperature 
digenite (stable above 60° C) was found 
by these workers to have a cubic symmetry. 
The present study was undertaken to ob- 
tain much-needed crystallographic infor- 
mation about the high-temperature form 
of digenite and bornite, as well as of the 
continuous solid solution series existing be- 
tween these two minerals at elevated tem- 

Natural Bornite 

Single crystals of bornite from the Cop- 
per Corp. Mine, Ontario, Canada, were 
studied by Donnay, Donnay, and Kullerud 
(Year Book 57, pp. 248-249). During this 
last year bornite single crystals from the 
locality were heated for 8 months at 150° C 
in evacuated silica tubes and studied in pol- 
ished sections as well as by crystallographic 
methods. An inversion in natural bornite 
at about 170° to 200° C was reported by 
Frueh (1950) and was also reported to 
occur in synthetic bornite at 190° C by Kul- 

lerud and Roseboom (1958). Bornite single 
crystals that were heated at 150° C for 
8 months were found by precession and 
Weissenberg photographs to possess a te- 
tragonal symmetry with a = 10.94 ±0.02 A 
and c=21.85±0.04 A. The space group 
was determined without ambiguity to be 
D 2 (i 4 -P42ic from the extinction rules of the 

Synthetic Bornite 

Single crystals of CusFeS* were synthe- 
sized from the elements in the appropriate 
proportions in evacuated silica tubes at 
900° C. The material was first heated at 
700° C until all sulfur combined with the 
metals; then the tube was opened under 
acetone and the material was finely 
ground. This mixture was inserted in an- 
other silica tube, which again was evacu- 
ated and heated at 900° C for 96 hours and 
then quenched directly in water. Through 
the silica tube wall the quenched material 
appeared homogeneous and of a yellow, 
brassy color. When the tube was opened 
this material reacted immediately with the 
atmosphere, and its surface took on a blue 
iridescent color identical to that seen on 
natural bornite. Fresh surfaces developed 
by grinding under acetone showed the 
original brassy color seen through the tube 

Several single crystals were examined by 
means of rotation, Weissenberg, and pre- 
cession X-ray methods. Two forms of born- 
ite were observed. One was cubic with the 
diffraction aspect F*** and the cell edge 
10.97 ±0.02 A. The other form was cubic 
with primitive lattice and cell edge 21.90 ± 
0.05 A. Powder patterns of the two forms 
appear identical on chart or film because 
the reflections distinguishing the two forms 
from each other are too weak to be seen. 
Although the diffraction pattern of the 
crystal with primitive lattice is complex, it 
can be explained as due to twinning of the 
tetragonal form described above. The twin 
axis is [221]. It seems, therefore, that the 
apparent cubic bornite with cell edge 
21.90 A is a quench product. The real 



high-temperature bornite is a cubic form 
with a=10.97 A, as first described by Lund- 
quist and Westgren (1936), but with dif- 
ferent diffraction aspect F***. 


It appears that bornite has two poly- 
morphs; the high-temperature form is 
cubic (a— 10.97 A, F***), and the low- 
temperature form is tetragonal (a= 10.94 ± 
0.02 A^_ c=21.85± 0.048 A, space group 
D 2 <z 4 -P42iC ) . The relation between the high 
and low forms is similar to that between 
the high and low forms of chalcopyrite, ex- 
cept that the cell dimensions of bornite are 

TABLE 30. Comparison of the Crystallographic 
Data of Chalcopyrite and Bornite 

Low Form High Form 

Chalcopyrite D,J--iMd F*** 

CuFeS 2 Z=4 Z=2 

(Donnay, Donnay, 
and Kullerud, a=5.24 A «=5.264 A 

Year Book 57) ^=10.34 A 

Bornite D^-FAl^c F*** 

Cu 5 FeS 4 Z—\6 Z=8 

a=10.94A a=\0.97 A 
c=21.85 A 

twice those of chalcopyrite. These relations 
are shown in table 30. 

Four types of natural bornite have been 
described so far in the literature: (1) Cubic 
type with a =10.93 A (Lundquist and 
Westgren, 1936). (2) Orthorhombic type 
of pseudotetragonal cell with a — b — 21.90 
A, c- 10.95 A (Frueh, 1950; Donnay, Don- 
nay, and Kullerud, Year Book 57). (3) 
Cubic type with Fd**, a = 21.94 A (Don- 
nay, Donnay, and Kullerud, Year Book 
57). (4) Cubic type with a = 32.8 A (Tun- 
ell and Adams, 1949). 

No satisfactory explanation has been 
given to clarify the relations among these 
four types in addition to the tetragonal 
type described above, although the ortho- 
rhombic type might be explained by twin- 
ning of the tetragonal one. 

Our results confirm Frueh's (1950) de- 
scription of a cubic bornite with a = 10.97 A 
as a high-temperature form. 

Born ite-Digen ite Mix- Crystals 

Single crystals of three different compo- 
sitions on the bornite-digenite join have 
been synthesized (as described above), and 
examined by the precession and Weissen- 
berg methods. The compositions of the 
mix-crystals are: (1) Cu 2 4Fe 3 Si 7 (3 born- 
ite 4-1 digenite); (2) Cui 4 FeS 9 (1 born- 
ite 4-1 digenite); and (3) Cu 23 FeSi4 (1 
bornite 4- 2 digenite). 

Crystals of composition Cu 2 4Fe 3 Sn. Sin- 
gle crystals of this composition were syn- 
thesized at 600° and 700° C. The 600° and 
700° C crystals gave identical patterns 
when studied by single-crystal X-ray meth- 
ods, showing cubic symmetry with a — 
21.98 ±0.02 A and diffraction aspect F***. 
Only the reflections (8m ±L, Sn±L, L) 
appear, where m and n are integers. 

Crystals of composition CuuFeSg. Single 
crystals of this composition were synthe- 
sized at 700° C and then heated at 400° C 
for 6 months. They are cubic with a = 
22.21 ±0.02 A. The diffraction aspect is 
F** # , as for the Cu 2 4Fe 3 Si7 composition 
crystals, but the intensity distribution is 
slightly different from that of the latter 
crystals. Only the reflections ($m±L, 
Sn ± L, L) appear. 

Crystals of composition Cu 23 FeSn. Sin- 
gle crystals of this composition were syn- 
thesized at 700° C. They are cubic with 
a = 27.69 ±0.02 A and the diffraction aspect 
F***. Only the reflections of (10m ±L, 
10«±L, L) appear. These results are in 
agreement with those reported for pure 
digenite (CugS s ) by Donnay, Donnay, and 
Kullerud (1958). 

Conclusions. The cubic high-tempera- 
ture form of the Cu 9 S 5 -Cu 5 FeS4 solid solu- 
tion series cannot be quenched for all com- 
positions even by rapid cooling. The X-ray 
properties of single crystals of intermediate 
composition showing only 8m ±L, $n±L, 
L reflections when examined at room tem- 


perature can be explained on the basis of 
twinning of the low-temperature form 

similar to that described in digenite by 
Donnay, Donnay, and Kullerud (1958). 


F. Chayes 

The Basic Computation 

Interest in the diffraction effects of dis- 
order and short-range ordering was ini- 
tially aroused by the casual discovery that 
the reported spacing of certain subsidiary 
reflections in the diffraction patterns of 
some intermediate plagioclases was directly 
proportional — with proportionality con- 
stant of unity to the limits of experimental 
error — to the proportion of Al atoms in 
the silicate framework. In direct space the 
parameter of interest would be the recipro- 
cal of this quantity, or of 1 , where a = 

Now it happens that a" 1 is the expected 
value of the number of consecutive sites 
occupied by Si in a random sequence con- 
taining N&i Al atoms and A/si Si atoms. 
The coincidence suggested the puzzling 
possibility that these subsidiary reflections, 
ordinarily interpreted as ordering effects, 
might in fact be indicative of disorder. 
Though it did not seem at all likely that 
such a possibility had been overlooked by 
specialists in the field, the theory about the 
generation of subsidiary reflections in inter- 
mediate plagioclase was (and remains) in 
a most unsatisfactory state. 

It seemed desirable to begin by deter- 
mining whether variation in short-range 
ordering as controlled by numbers of runs 
generated any systematic and appreciable 
diffraction effects. For this purpose a very 
simple model was chosen, in which each 
element is a row of identical, identically 
spaced, points, or, in a diffraction mask, 
holes. The distance between rows is the 
same throughout, and adjacent rows are 
either in exact register or offset from each 
other by some fraction (t) of the distance 
between holes. If we take Y normal and X 
parallel to the row direction in an array 
containing N rows, the coordinates of the 

first point in the nth. row are X n — or t, 
depending on whether the row is in reg- 
ister with or offset from the first row, and 
Y n =(n-\)/N. 

The testing was originally to be by di- 
rect optical experimentation, and it was 
expected that large numbers of masks 
would be required. The model described 
above greatly simplifies the design and pro- 
duction of such masks; the same properties 
that make it so desirable for this purpose 
subsequently proved even more convenient 
in the high-speed computation that re- 
placed optical experimentation shortly after 
publication of last year's report. 

The results of the early experimentation 
were highly promising; well defined dif- 
fuse reflections were in evidence, and a 
striking effect, a doubling of the most 
prominent diffuse reflection from short- 
range ordered arrays, became even more 
pronounced with early increase in mask 
length and improvement of the diffrac- 

A considerable further extension of mask 
length seemed desirable but could be ac- 
complished only by building a more elab- 
orate mask generator; the viewing of such 
masks would in turn have required a more 
elaborate and much larger diffractometer. 
It seemed sensible to examine the feasibility 
of shifting from optical experimentation, 
essentially a form of analog computation, 
to high-speed digital computation, and un- 
der the circumstances it was natural to use 
the diffractometer, rather than a real or 
imagined crystal, as a model; the first prob- 
lem was to develop a code that would cal- 
culate, from a given array, an intensity 
profile like that generated by the diffrac- 
tometer from a mask constructed from the 
same array. 

For light from a point source passed 


through an array of infinitely small holes 
distributed in the fashion described above, 
the intensity of the diffracted beam of 
order (hJ() is proportional to 




Z sin 2rr(hx n + \y n ) 
L 1 


Z cos 2n(hx n + \yn) 


a summation that can readily be performed 
by digital computation. We desire a pro- 
file along \ for an arbitrary h, and 0<^ 
<N; the time needed for the calculation 
will be directly proportional to N and the 
number of hk values required, since each 
hk calls for N sines and cosines if equa- 
tion 1 is to be used directly. For N = 1000, 
the contemplated length, and 0<^(Y 2 ) 
<N, for instance, a single calculation 
would require 2X 10 6 sines and cosines; the 
time needed for the rest of the calculation 
is trivial, but generating these sines and 
cosines would take nearly 4 hours on a ma- 
chine of the capacity of the IBM 704. This 
is perhaps not prohibitive where only a few 
computations are to be run. It was sus- 
pected from the outset, however, that a 
large number of computations would be 
needed, so that unless drastic economies 
could be effected without significant sac- 
rifice of accuracy the substitution of nu- 
merical for optical experimentation would 
be utterly impracticable. It is precisely here 
that the properties of the rovv-by-row array 
proved most useful. It is tedious but not 
difficult to show that for an array of this 
type equation 1 may be put in the form 

N N 

where S — Z sin 2n^y», C = Z cos Inkjn, 

1 1 

K = 2nky, and subscript t indicates that 
summation is to extend only over the Nt 
points for which X = t, i.e. for rows that are 

Initially it was hoped that a rather broad 
spacing of \ would be satisfactory, but the 
first few computations suggested that all 
integral ^'s would be needed and it was 
finally decided that hk for \ an odd multi- 
ple of ! /2 would also be helpful. The neces- 
sary sines and cosines for the latter are 
readily obtained from those for the next 
succeeding integral ^'s. Equation 2 thus 
generates four intensities from each set of 
Nt sines and cosines as compared with one 
generated by equation 1 from each set of 
N; the resulting reduction in machine time 
is more than eightfold, since we may al- 
ways take Nt<N . 

Finally, if the sines and cosines are 
grouped into subsums depending on 
whether, for integral a, (#— T)— 4tf=0, 1, 
2, or 3, all the necessary sums of sines and 
cosines for intensities in the range 0<^ 
( 1 / 4)<iV can be found from those calcu- 
lated over 0<^(l)<2V/8. The net result 
of these involved substitutions is rather 
startling. If we suppose that in our 1000 
"layer" mask Nt = 250, iV o = 750, for in- 
stance, equation 2 yields from only 33,000 
sine-cosine pairs the same information ob- 
tained from equation 1 with 2 million such 
pairs. The generation of 33,000 sine-cosine 
pairs is a matter of less than 4 minutes' 
machine time, so that, if large numbers of 
masks were indeed required, the work 

I hk = S 2 + C 2 + 2(1- cos 2nht) (LsmKt) + (LcosKtJ 


Z sin Kt 

- N N 

S L sin Kt+CL cos K 

- l l 


+ 2 sin 2nht S Z cos Kt — CT. sin Kt 

■2(l — cos2nht) 



h<.N-k)=hie—4 sin 2nht 

F N 

l s 7 

cos Kt 






could probably be done at far less expense 
by numerical than by optical experimenta- 

At about this stage in the proceedings the 
project sustained two rather severe shocks. 
Professor Johannes Burgers, of the Univer- 
sity of Maryland, and Dr. M. G. Bown, of 
Oxford University, showed independently 
that the model chosen was not one from 
which discrete subsidiary reflections could 
be expected. Further, the computed trans- 
forms bore little if any resemblance to 
those obtained by diffractometer from the 
same sequences. These difficulties are dis- 
cussed below in order. 

Diffuse Diffraction from the Row-by-Row 

Dr. Bown's interest was stimulated by 
the preliminary work reported in Year 
Book 56. In this, as in Year Book 57, the 
illustrative material was entirely concerned 
with masks in which the offset between 
rows was half of the distance between holes 
in a row (t= l / 2 ). For this arrangement he 
obtained the diffuse component of diffrac- 
tion both from an interatomic vector map 
and by an extension of Wilson's (1949) 
theory of "mistakes" in layered structures. 

Professor Burgers solved the general 
case (0<O<l), finding that the intensity 
of diffuse scatter of order (h\) is propor- 
tional to 


_ 2o(l-o)(l-cosY)(l-r 2 ) 
1 — 2r cos 8 + r~ 


where o = N /(N +N t ). 

r= l — p — q;p = Pr(xj+i = 0\xj= 
t); q=Pr(xj+i = t\xj = 0). 

It is clear from equation 3, of which 
Bown's solution for t =l / 2 is a special case, 
that the intensity of the diffuse diffraction 
varies continuously, symmetrically, and 
rather gradually about a single maximum 
or minimum at t{=N/2. For arrays that 
depart from disorder in the direction of 
short-range ordering, r<0 and I'un/2 is a 

maximum; at complete disorder, r—0 and 
the diffuse diffraction is a streak of uni- 
form intensity; in an array containing 
fewer runs than characteristic of disorder, 
i.e. a dissociated or "unmixed" array, r>0 
and I'hN/2 is a minimum, o and t operate 
entirely as scaling factors affecting the 
absolute value of /'; only changes in r, i.e., 
the level of ordering, can change the shape 
of the curve. And no permissible variation 
in any of the parameters can introduce 
either discontinuities or additional maxima. 
The model is quite evidently not one suita- 
ble for optical or numerical experimenta- 
tion aimed at examining the relation 
between short-range ordering and the loca- 
tions of discrete subsidiary maxima. 

The derivation involves geometric sum- 
mations in which N is formally infinite, 
but this is not a serious restriction. N need 
only be large enough that r j — >0 for 
/'<<iV, and by definition |r|<l; equation 
3 should be applicable for values of N 
within experimental reach. The optical 
transforms ought not to contain subsidiary 
maxima except at \=N /2. But they do, 
and the only reasonable explanation is that, 
although a satisfactorily large N was per- 
haps within experimental reach, it was not 
in fact reached. (Some elegant photographs 
published by Willis, 1958, indicate that, for 
a^Y 2 and certain values of r, TV =1000 
is large enough. In my optical masks 

Comparison of Computed and Observed 

The basic computation yields an exceed- 
ingly jagged intensity profile, quite unlike 
anything actually observed in the diffrac- 
tometer. It assumes a point source and a 
mask in which point scatterers are repre- 
sented by infinitely small holes; in such a 
system there would be no energy loss and 
no limitation on resolving power except 
that inherent in the lens system. 

The size of the openings in the mask is 
principally concerned with the fall-off of 
intensity from / (0 ,o) outward. It is of no 



immediate concern unless intensities are 
to be measured photometrically, either on 
the photographed transform or directly on 
its image at the focal plane of the diffrac- 
tometer. The latter procedure has been 
used by Willis (1958) ; the measured inten- 
sities must then be corrected by an Airy 
modulation before comparison with theo- 
retical expectation. Since there is no fall-off 
in the basic computation, we have no need 
for such a correction in the calculated 

The size of the entrance aperture, on the 
other hand, is of critical importance. In 
the diffractometer considered as a model 
of the experimental situation in X-ray 
analysis, the diameter of the entrance aper- 
ture plays the role of crystal dimension. 
Unfortunately, however, it plays according 
to a very different set of rules, and in this 
respect the analogy between X-ray and op- 
tical arrangements is not close. 

In X-ray crystallography the collimated 
incident beam is usually much broader 
than the crystal, the only "entrance aper- 
ture" in the system is the crystal itself, and 
the wavelength of the incident radiation is 
very small in relation to this "aperture." 
The relation between crystal size and re- 
flection size is reciprocal, and the intensity 
of the incident beam may be taken as 

In the diffractometer, collimation is ob- 
tained by placing an entrance aperture at 
the focal plane of the collimating lens, the 
transform is focused by a collecting lens, 
which, in the most efficient arrangement, is 
of the same focal length as the collimator, 
and the incident wavelength is large 
enough in relation to aperture diameter to 
introduce considerable variation in inten- 
sity across the collimated beam. The result 
is that if the mask occupies the entire area 
of the collector lens each reflection of order 
{hJ() is a direct 1: 1 image of the aperture 
with intensity distribution like that across 
the aperture, i.e. dropping off abruptly 
from a rather flat central maximum. If, as 
is usual, the mask is noncircular and occu- 
pies less than the full area of the collector 

lens, each reflection of order (hf{) will be 
a distorted and somewhat enlarged image 
of the entrance aperture, the enlargement 
varying inversely with the ratio of mask 
dimension to lens diameter. 

The ratio of aperture diameter to trans- 
lation distance between principal reflections 
at the focal plane of the collector is the 
major control on the resolving power of 
the instrument. Its effect may be modeled 
numerically by passing through the inten- 
sities given by the basic computation a 
moving average of width w = Nzd/c*; 
where N is the number of rows in the 
mask (and also of orders per period), z is 
the number of evenly spaced calculated in- 
tensities per order, d is the diameter of the 
entrance aperture, and c* is the distance 
between ho and IhN at the focal plane of 
the collector. 

The variation of intensity across the 
aperture calls for an average weighted so 
that observations distant from the center 
by more than ±#//4 contribute very little 
to the result; in this respect an unweighted 
average of width w materially underesti- 
mates the resolving power of the instru- 
ment. By enlarging the image of the aper- 
ture, on the other hand, masks that do not 
utilize the full diameter of the collector 
lens produce an opposite effect. Although 
weights have been developed, both by cal- 
culation and by experiment, they have not 
yet been incorporated into the machine 
program. (In view of the results of the 
preceding section such a refinement would 
appear to be something of a tour de force.) 

The effect of a properly weighted mov- 
ing average can be roughly approximated, 
however, by varying the width of an un- 
weighted moving average. In the arrange- 
ment used to generate the transform in fig- 
ure 45, plate 2, Year Book 57, for instance, 
w~15. The ratio of lens diameter to mask 
length is 13/5, so that an individual reflec- 
tion should be of width 15(13/5) =39 
observations. Recalling that the intensity 
profile across such an image is rather flat- 
topped over the central half and then drops 
off precipitously, we approximate the over- 


all effect by simply discarding the outlying 
observations and using an unweighted 
moving average of 19 observations (or 21, 
the present code requiring an odd num- 
ber!). The result is shown in figure 63. 
The profile is rather jagged, but the prin- 
cipal features and most of the detail of the 
optical transform are retained. In particu- 
lar, the doubling of the central node is as 
clear in the calculated transform as in the 
optical one. 

cient length is evidently a biased but per- 
haps also a consistent estimator of ^max. If 
so, an average across a group of such arrays 
ought to converge, in a probability sense, 
on an intensity distribution with a single 
central maximum, as predicted by equa- 
tion 1; the larger the number of groups in 
the array, the smaller the expected differ- 
ence (^average max — A//2). But, no matter 
how many arrays in the group, the average 
value of the quantity (^max — A//2) =7^0; 

Fig. 63. Calculated distribution of diffuse intensity generated by a short-range ordered "layered" 
sequence, N= 192, N t = 64, using moving average of 19 "observations" or 9/4 orders. (The array 
is the same as that used to generate the optical transform shown in Year Book 57, p. 243, fig. 45, 
pi. 2.) 

Averaging within and across Arrays 

We are now in a position to examine a 
little more closely the discrepancy between 
the result expected from equation 1 and 
that actually found, shown in figure 63. We 
may regard an array such as that which 
yields figure 63 as a sample of a very large 
population of arrays of the same length, 
composition, and level of ordering, any one 
of which provides an estimate of the order 
at which the diffuse intensity reaches a 
maximum. Equation 1 guarantees that if 
N is sufficiently large the maximum will 
be centered at ^max=A//2. 

In the experimental work so far reported 
t=Y 2 , and the diffuse diffraction is con- 
fined to zones for which h is an odd in- 
teger. For these values of h and t, equa- 
tion 2b shows that the diffuse diffraction 
will be symmetrical about \ = N/2 regard- 
less of the size of A/. 12 An array of insuffi- 

12 I believe the symmetry relation is much more 
general but have been unable to find a satis- 

the average intensity 
group of 15 arrays of 

this quantity is an estimate of the bias, and 
an increase in the number of arrays on 
which it is based merely improves the 

Figure 64 shows 
distribution for a 

length N— 192. Each array is a replicate of 
the one whose transform is shown in fig- 
ure 63, except that the order of runs is sepa- 
rately randomized. In each, A7t = 64, N = 
128, and the relative frequency of runs of 
lengths 1, 2, 3, • • •, is that to be expected 
at perfect short-range order. The open 
circles are calculated from equation 1, and 
the agreement is obvious. But, although 
there is almost no indication of a doubling 
of the central node in the average intensi- 
ties, the average value of (^max — A//2) is 
4.4; in the transform of an individual array 
of this length one would expect, instead of 

factory proof that does not require that N in- 
crease without limit. It is certainly true for any 
integral ^, for in this case S = C = 0, and the 
second term of (2b) vanishes. 



Fig. 64. Within-and-across average of 15 sequences like that used to obtain figure 63. 

a single central maximum at \—96, a 
doublet with maxima at ^i=91.6, \o — 
100.4. This expectation would of course be 
subject to a rather generous sampling 

Averaging across Arrays 

One implication of the preceding discus- 
sion is that, if enough arrays are available 
for averaging and there is no need to 
model the diffractometer, the moving aver- 
age within arrays may be dispensed with. 
Figure 65 shows the result of such a calcu- 
lation for the same set of arrays used to 
obtain figure 64. Comparison with figure 
64 shows the efficiency of the moving aver- 
age as a means of damping high-frequency 
oscillations. This is, of course, one of the 
principal functions of the entrance aper- 
ture in the difTractometer. 


The relations between short-range order- 
ing and subsidiary reflections remain essen- 
tially untested by this work. The incon- 
clusive results so far obtained are in no 
way the responsibility of the numerous per- 
sons who have helped me get them. My 
indebtedness to J. S. Burgers and M. G. 
Bown will be obvious to readers of the 
second section of this report; to the former 
I am particularly grateful for many hours 
of helpful discussion. J. M. Cameron and 
Gordon J. F. MacDonald coded and ran 
the first computations, and finally suc- 
ceeded in teaching me how to do my own 
programing. Dr. Isabella Karle kindly 
prepared microdensitometer traces from a 
number of negatives. Most of the comput- 
ing was done at the National Bureau of 

12 24 36 48 60 72 84 96 108 120 132 144 156 168 180 192 

Fig. 65. Across-array average of same 15 sequences used to obtain figure 64. 



Penologists' Club 

Seven meetings of the Petrologists' Club 
were held at the Laboratory this year. 
Average attendance was eighty-five. A field 
trip to examine the geologic setting of the 
Baltimore gneiss in the vicinity of Balti- 
more was led by C. A. Hopson, of The 
Johns Hopkins University. W. G. Ernst 
was elected to serve as co-chairman with 
D. B. Stewart for 1960. The following 
papers were presented: 

"Crystallization of feldspar-silica mixtures 
from liquid," by D. B. Stewart (U. S. Geo- 
logical Survey). 

"The influence of the oxidation state of iron 
on the mineral assemblages of some Scottish 
metamorphic rocks," by G. A. Chinner (Geo- 
physical Laboratory). 

"The metamorphic complex at Cooma, N. 
S. W., Australia," by G. Joplin (Australian 
National University). 

"Carbonate deposition in open bodies of 
water," by E-an Zen (University of North 

"Phase relations of biotites on the join 
phlogopite-annite," by D. R. Wones (Geo- 
physical Laboratory). 

"Synthetic carbonatite magmas," by P. J. 
Wyllie and O. F. Tuttle (Pennsylvania State 

"Deposition of sulfide ores from sulfide- 
deficient solutions," by T. S. Lovering (U. S. 
Geological Survey). 


The Laboratory continued its weekly 
series of seminars, with papers presented 
largely by staff members and concerned 
mainly with discussions of work in prog- 
ress. The following talks were given by 
guest speakers from outside the Laboratory : 

"Growth of crystals from vapor," by R. S. 
Bradley (Leeds University, England). 

"Variation of length of day during histori- 
cal times," by Gordon J. F. MacDonald (In- 
stitute of Geophysics, University of California, 
Los Angeles). 

"Uranium ore deposits," by Nobuo Kata- 
yama (Tokyo University). 

"Paramagnetic resonance in asphaltenes," 

by Robert Rex (California Research Corpora- 

"Earth's crust and the origin of life," by 
J. D. Bernal (London University). 

"Mohole Project," by Willard Bascom (Na- 
tional Academy of Sciences). 

"Lattice defects, dislocations, and meta- 
morphism," by Johannes Burgers (Univer- 
sity of Maryland). 

"Problems of Sudbury geology," by T. C. 
Phemister (University of Aberdeen). 

"Role of oxygen pressure in the crystalliza- 
tion and differentiation of basaltic magma," 
by E. F. Osborn (Pennsylvania State Univer- 

An informal seminar on the general sub- 
ject "Thermodynamics of Open Systems" 
was held at the Geophysical Laboratory on 
November 17, 1958. Principal speakers 
were Dimitri S. Korzhinsky, of the Insti- 
tute of Geological Sciences, Academy of 
Sciences, USSR; James B. Thompson and 
Robert M. Garrels, of Harvard University. 
The speakers, Laboratory staff, and guests 
participated in a stimulating all-day dis- 
cussion. Guests included E-an Zen, of the 
University of North Carolina; H. P. Eug- 
ster, of The Johns Hopkins University; 
E. T. Roedder, P. B. Barton, W. T. Pecora, 
and D. B. Stewart, of the U. S. Geological 

Fellows of the Geophysical Laboratory 
gave a series of seminars at the Department 
of Geology, The Johns Hopkins Univer- 
sity, as follows: 

"Stability and phase relations of alkali am- 
phiboles," by W. G. Ernst. 

"Phase relations in the Fe-As-S system and 
applications to natural assemblages," by L. A. 

"The system Cu-Fe-S and its application 
to copper deposits," by E. H. Roseboom. 

"Pyrrhotite as a geologic thermometer," 
by R. G. Arnold. 

"Stability relations of a nonaluminous bio- 
tite," by D. R. Wones. 

"Subsolidus relations in the system FeO- 
Fe 2 3 -Al 2 3 ," by A. C. Turnock. 

"Thermodynamic limitations on low-tem- 
perature ore solutions," by H. L. Barnes. 


"Metamorphic mineral assemblages from 
Glen Clova, Scotland," by G. A. Chinner. 

"Compositions of coexisting feldspars and 
muscovite and the crystallization of pegma- 
tites," by P. M. Orville. 

"Stability of cordierite," by W. F. Schreyer. 

"The Ni-As-S system," by R. A. Yund. 


During the report year staff members 
were invited to present lectures as follows: 

P. H. Abelson lectured at the Washing- 
ton Academy of Sciences, and at a sym- 
posium on chemical approaches to the rec- 
ognition of petroleum source rocks held 
during the Fifth World Petroleum Con- 
gress. As a participant in the American 
Geological Institute Visiting Geoscientist 
Program, he gave a series of lectures at the 
departments of geology at Lafayette Col- 
lege and Lehigh University. 

H. L. Barnes addressed the Houston 
Mineralogical and Geochemical Club at 
Rice Institute. 

F. Chayes lectured at the Washington 
Crystal Colloquium. 

S. P. Clark, Jr., gave two lectures at the 
Department of Geology of the University 
of Chicago, as well as addressing the Ex- 
plosions Research Department Colloquium 
at the Naval Ordnance Laboratory and the 
Geological Society of Washington. 

G. Donnay lectured at the Universities 
of Liege and Ghent, Belgium; Zurich, 
Switzerland; Paris, Grenoble, and Bor- 
deaux, France. She also addressed the 
Fritz-Haber-Institut der Max-Planck-Ge- 
sellschaft in Berlin-Dahlem and gave an 
invited paper at the Fedorov Session on 
Crystallography, which was held in Lenin- 
grad on May 21-27, 1959. 

W. G. Ernst gave a series of lectures at 
the Department of Geology, University of 
California at Los Angeles. He also ad- 
dressed the Earth Science Colloquium and 
the Department of Geology at the Massa- 
chusetts Institute of Technology, and the 
Geological Society of Washington. 

T. C. Hoering lectured at the Depart- 
ment of Chemistry, Fordham University; 
a symposium on isotope geology sponsored 

by the American Petroleum Institute at 
Tulsa, Oklahoma; and the Geological So- 
ciety of Washington. 

G. Kullerud gave six lectures at the 
Summer Institute in Geology for college 
teachers at the University of Illinois and 
two lectures at Lehigh University, Bethle- 
hem, Pa. He addressed the New York Min- 
eralogical Club at Columbia University. 

J. F. Schairer gave a series of lectures to 
the chemistry and geology classes at St. 
John's College. He also served as a panel- 
ist at the Cleveland meeting of the Ameri- 
can Institute of Mining and Metallurgical 

G. R. Tilton lectured at the Department 
of Geology, Columbia University. 

H. S. Yoder, Jr., gave three lectures at 
the Summer Institute in Geology for col- 
lege teachers at the University of Illinois 
and addressed the Department of Geology 
and Institute of Geophysics at the Univer- 
sity of California at Los Angeles. He 
served as visiting Professor of Geochemis- 
try for the fall term, 1958, at the California 
Institute of Technology. Lectures on Ad- 
vanced Problems in Petrology were shared 
with Professor C. E. Tilley. 

The "Summary of Published Work" be- 
low briefly describes the papers published 
in scientific journals during the report year. 
In addition, the following papers are now 
prepared for publication: H. L. Barnes, 
"The effect of metamorphism on metal 
distribution near base metal deposits"; 
F. R. Boyd, "Geology of the Yellowstone 
rhyolite plateau"; S. P. Clark, Jr., "Effect 
of pressure on the melting points of eight 
alkali halides"; E. D. Eanes and G. Don- 
nay, "Dimerization of trans-c'mnamic to 
a-truxillic acid"; W. G. Ernst, "The sta- 
bility relations of magnesioriebeckite"; 
W. F. Libby, "Radioactive fallout particu- 
larly from the Russian October series"; 
P. M. Orville, "Composition and geology 
of several unzoned pegmatites in the Key- 
stone District, Black Hills, South Dakota"; 
E. G. Zies and F. Chayes, "Pseudoleucite 
in a tinguaite from the Bearpaw Moun- 
tains, Montana." 




(1282) Isotopic composition of lead from tek- 
tites. G. R. Tilton. Geochim. et Cos- 
mochim. Acta, 14, 323-330 (1958). 

The isotopic composition of lead in three 
tektites and Libyan Desert glass is compared 
with that in known terrestrial and extraterres- 
trial sources. The lead contained in the 
glasses is similar to modern terrestrial lead, 
particularly lead from modern oceanic sedi- 
ments. The uranium, thorium, and lead con- 
centrations were determined for one of the 
glasses, an australite. Evidence is given which 
indicates that within the last tens of millions 
of years differentiation of uranium, thorium, 
and lead occurred in the parent material of 
the australite. These results are difficult to 
explain in terms of any extraterrestrial origin 
involving fusion of materials from the moon, 
meteorites, or comets, but they are readily ex- 
plained if tektites are of terrestrial origin in- 
volving fusion of argillaceous sediments in 
some unspecified way. 

(1284) Experimental studies on micas. A syn- 
thesis. H. S. Yoder, Jr. Proc. Sixth 
Natl. Conf. on Clays and Clay Minerals, 
42-60 (1959). 

The principal end members of the micas 
believed to be common in sediments have been 
synthesized and some of their stability rela- 
tions determined. The polymorphs of mus- 
covite and paragonite, the principal diocta- 
hedral end members, obtained were lMd, 1M, 
and 2Mi, and those of phlogopite and annite, 
the principal trioctahedral end members, lMd 
and 1M or 3T. The range of stability of each 
of the polymorphs could not be fixed ac- 
curately because of the slow rate of transfor- 
mation; however, the transformations lMd — > 
1M— »2M! were effected for muscovite and 
paragonite and 1M— ■► lMd or 3T and 2Mn 
— > 1M or 3T for phlogopite. The growth 
characteristics of these micas in the labora- 
tory are believed to be analogous to the for- 
mation of micas in sediments. 

Knowledge of the synthetic micas con- 
tributes greatly to an understanding of the 
natural materials called illite, hydromica, and 
high-silica sericite. The dioctahedral mem- 
bers of these materials and related minerals 
may be delineated accurately in the system 
muscovite - Al celadonite - pyrophyllite and 

their iron analogues. The trioctahedral mem- 
bers of some of the same materials may be 
outlined in the system phlogopite-eastonite- 
talc and their iron analogues. The postulated 
substitution schemes in these systems are 
mainly MgSi ^± A1 VI A1 IV , KA1 ^± Si, and 
H 3 ^± K. In materials intermediate between 
these systems, such as most biotites and ver- 
miculites, the substitution of 3Mg ^± 2A1 VI 
is of major importance. The mixed-layer 
structures involving micas are elucidated. 

(1285) Radioactive ages of micas from granitic 
rocks by Rb-Sr and K-A methods. L. T. 
Aldrich, G. W. Wetherill, G. L. Davis, 
and G. R. Tilton. Trans. Am. Geo- 
phys. Union, 39, 1124-1134 (1958). 

Analytical data and ages are given for K-A 
Rb-Sr, U-Pb, and Th-Pb ages from a number 
of locations. Comparisons of the ages ob- 
tained indicate (a) that K-A and Rb-Sr ages 
of mica usually agree when modern labora- 
tory-determined values of the decay constants 
of K 40 and Rb 87 are used; (b) that the K-A 
and Rb-Sr ages of micas agree with the U-Pb 
ages of cogenetic uraninite, zircon, and mona- 
zite when the U-Pb ages are concordant; (c) 
that, when the U-Pb ages of these uranium 
minerals are discordant, the mica ages agree 
best with the Pb 207 -Pb 206 age; (d) that the 
pattern of K-A and Rb-Sr ages of micas can 
be most simply explained by postulating the 
loss of variable but small amounts of argon 
from the mineral. 

(1286) Ages of minerals from the Baltimore 
gneiss near Baltimore, Maryland. G. R. 
Tilton, G. W. Wetherill, G. L. Davis, 
and C. A. Hopson. Bull. Geo!. Soc. 
Am., 69, 1469-1474 (1958). 

Biotite from the Baltimore gneiss, consid- 
ered to be part of the Precambrian basement, 
was found to be 310 million years old by the 
Rb-Sr method. Microcline by the same 
method and zircon by U-Pb and Th-Pb meas- 
urements are much older, nearly 1100 million 
years old. The results are interpreted as im- 
plying the existence of a 1050-million-year-old 
basement that was metamorphosed during 
the Paleozoic era. 

Detailed rock descriptions are given, as well 
as age data for Precambrian gneisses not 
affected by subsequent metamorphism. 


(1287) Natural and synthetic ferroselite: A 
roentgenographic mimesis of rammels- 
bergite. G. Kullerud and G. Donnay. 
Geochim. et Cosmocfiim. Acta, 15, 73- 
79 (1958). 

Iron diselenide (FeSe 2 ) has been synthe- 
sized in sealed, evacuated silica glass tubes at 
temperatures between 450° and 550° C. Its 
powder pattern is identical with that of fer- 
roselite described by Buryanova and Komkov 
(1955). The first 15 lines of the pattern are 
indistinguishable from those of rammelsberg- 
ite (NiAs 2 ). Single crystals of ferroselite from 
Temple Mountain, Utah, were studied on the 
precession camera. Their diffraction patterns 
are also deceivingly similar to those of ram- 

(1288) Origin of granite in the light of experi- 
mental studies in the system NaAlSi 3 O g - 
KAlSi 3 8 -Si0 2 -H 2 0. O. F. Tuttle and 
N. L. Bowen. Geol. Soc. Am. Mem. 74, 
xi + 153 pp., 67 figs., 6 pi. (1958). 

This paper deals with the experimental de- 
terminations of phase-equilibrium relations 
in the system NaAlSi 3 8 (albite)-KAlSi 3 8 
(orthoclase)-Si0 2 -H 2 and with the applica- 
tion of these results to some petrologic prob- 

The laboratory experiments can be di- 
vided into two categories: (1) study of the 
liquidus, solidus, and subsolidus phase rela- 
tions in synthetic mixtures; and (2) study of 
natural minerals and rocks. Phase relations in 
the binary systems Si0 2 -H 2 0, NaAlSi 3 O s - 
H 2 0, KAlSi 3 8 -H 2 0; the ternary systems 
NaAlSi 3 8 -KAlSi 3 8 -H 2 0, NaAlSi 3 8 - 
Si0 2 -H 2 0, KAlSi 3 8 -Si0 2 -H 2 0; and the 
quaternary system NaAlSi 3 8 -KAlSi 3 8 - 
Si0 2 -H 2 were investigated with the aid of 
synthetic mixtures. Analyzed natural feld- 
spars were used for subsolidus studies in the 
system NaAlSi 3 8 -KAlSi 3 8 , and the begin- 
ning-of-melting temperatures of natural gran- 
ites were studied and compared with those of 
synthetic granites. 

Phase studies in the quaternary system 
NaAlSi 3 8 -KAlSi 3 8 -Si0 2 -H 2 provided 
quantitative data on the melting relations in 
these granitic compositions as well as infor- 
mation on fractional and equilibrium crystal- 
lization. At constant pressure, the system is 
characterized by a minimum melting temper- 
ature on the boundary between quartz and 
feldspar solid solutions. Liquids throughout 

the system move toward this minimum on 
crystallization, and if fractionation is pro- 
nounced most liquids will reach the mini- 
mum. A plot of the normative albite, ortho- 
clase, and quartz in all analyzed granites and 
rhyolites from Washington's tables (1917) 
demonstrates that the minimum at low water- 
vapor pressure, corresponding to a water con- 
tent of 1 to 2 per cent, falls at the composition 
of the average granite and rhyolite. It is 
suggested that this demonstrates that crystal- 
liquid equilibria control granite compositions; 
therefore granites not formed at magmatic 
temperatures will be rare and will not have 
compositions related to the minimum. The 
liquids can originate by fractional crystalliza- 
tion of more basic liquids (i.e., basalts) or by 
fractional melting of appropriate sedimentary 
and metamorphic rocks. 

The beginning of melting of two granites — 
the Westerly, Rhode Island, and the Quincy, 
Mass. — one a normal calcalkaline granite, and 
the other an alkaline granite, was determined 
at a series of water-vapor pressures; a P-T 
curve for the beginning of melting of the two 
granites corresponds within the experimental 
error to the beginning of melting at the iso- 
baric minimum in the quaternary system. 

(1289) Annual report of the Director for 1957- 

(1290) Phase equilibria with particular refer- 
ence to silicate systems. J. F. Schairer. 
Chapter 5 in The Technique of Physico- 
Chemical Measurements at High Tem- 
peratures (ed. J. O'M. Bockris, J. D. 
Mackenzie, and J. L. White), London, 
Butterworths Scientific Publications, 

The method of quenching developed at the 
Geophysical Laboratory and employed in the 
study of phase-equilibrium relations in silicate 
systems of rock-forming oxides at high tem- 
peratures is described. The principal topics 
discussed are the preparation of homogeneous 
melts, crystallization of the melts and identifi- 
cation of the phases, the method of quench- 
ing, and the use of inert atmospheres with 
iron silicates. Complete details on the selec- 
tion of raw materials, preparation of charges, 
the melting furnace, crystallization methods, 
the quenching technique, the quenching fur- 
nace, regulation of temperature, measurement 
of temperature, calibration of thermoelements, 


and determination of equilibrium are given. 
The three principal methods (the Tuttle hy- 
drothermal quenching apparatus, the cold 
seal bomb, and the collapsible tube method) 
used for hydrothermal studies of silicate sys- 
tems at elevated temperatures are described, 
as well as the apparatus and techniques used 
for the study of silicate systems at high tem- 
peratures with high hydrostatic pressure. 

(1291) Crystal data on chlorophyll a. G. Don- 
nay. Arch. Biochem. Biophys., 80, 80- 
85 (1959). 

By using the hexagonal cell a — S.75 s , c = 
42.0 7 A, it is possible to index the eight sharp 
lines of the powder pattern of crystalline 
chlorophyll a. The cell volume, in view of the 
observed density of 1.079 g/cc, requires two 
molecules per cell. Chlorophyll is pyroelectric 
and optically active in solution, so that only 
space groups isomorphic with holoaxial point 
groups need be considered. In the hexagonal 
and orthorhombic systems molecular disorder 
must be assumed in any structural hypothe- 
sis. In the monoclinic system the only possi- 
ble space group is P2 with cell dimensions 
a = c = 8.75 8 , £ = 42.0 7 A, (3=120° 0'. Pack- 
ing considerations make this symmetry un- 
likely. The remaining possible space group 
is the triclinic PI in which the two molecules 
are not symmetrically related to each other. 
The only prediction that can be made about 
their packing rests on the presence of reflec- 
tion 1.10.0 — the planes of the chlorin rings 
should be parallel and should make an angle 
of about 29° with (010). 

(1292) The optical properties of heated plagio- 
clases. J. R. Smith. Am. Mineralogist, 
43, 1179-1194 (1958). 

The optical properties of nine chemically 
analyzed samples of natural plagioclases have 
been accurately determined before and after 
changing them to high-temperature modifica- 
tions by heat treatment. The change in Nx 
accompanying the structural change is slight 
in the composition range An to An 20 and 
negligible from An 20 to An 100. Ny and Nz 
change measurably in the composition range 
An to An 20, but negligibly in the re- 
mainder of the composition range. Measure- 
ments of principal refractive indices can there- 
fore give a reliable estimate of the composi- 
tion of a plagioclase regardless of its struc- 
tural state. 

The changes in optic axial angle accom- 
panying the structural changes are such that, 
given composition, measurements of optic 
axial angle serve to distinguish low-tempera- 
ture and high-temperature plagioclases in the 
composition ranges An to about An 40 and 
An 60 to about An 90. 

(1293) Symmetry of magnetic structures: Mag- 
netic structure of chalcopyrite. G. Don- 
nay, L. M. Corliss, J. D. H. Donnay, 
N. Elliott, and J. M. Hastings. Phys. 
Rev., 112, 1917-1923 (1958). 

The transformation properties of magnetic 
moments under symmetry and antisymmetry 
operations lead to 1421 possible space groups 
for ferromagnetic and antiferromagnetic crys- 
tal structures. A systematic procedure for 
magnetic structure determination is proposed, 
which takes into account the restrictions im- 
posed on spin directions by space groups. 
This method is applied to chalcopyrite, Cu- 
FeS 2 , which is found to be antiferromagnetic 
at room temperature: the chemical structure 
is that proposed by Pauling and Brockway, 
the space group M2d holds for the magnetic 
structure, in which the two iron (and possibly 
also the two copper) atoms tetrahedrally 
bonded to a common sulfur atom have anti- 
parallel spins directed along the c axis. A 
value of 3.85 Bohr magnetons per atom is 
found for the iron moment (0±0.20 p,B for 
copper). The possible existence of a second 
chalcopyrite modification in nature, suggested 
by conflicting results on material of Japanese 
origin, is ruled out, as specimens from both 
Ugo, Japan, and Joplin, Missouri, are found 
to have the same structure. 

(1294) Sine table for indexing powder patterns. 
J. D. H. Donnay and G. Donnay. Am. 
Mineralogist, 44, 177-179 (1959). 

Existing tables of natural values of the sine 
and cosine functions can be used to index 
powder patterns with greater accuracy than 
is provided by the specially constructed tables 
of interplanar spacings d or squares O of 
reciprocal-lattice vectors. In precision work 
20 O bs is read, and hence 29 C aic must be com- 
puted, to 0.01°. Available tables of d values 
and sin 2 for = 0(0.01 °)90° require inter- 
polation, for a step of 0.01° in corresponds 
to one of 0.02° in 20. By utilizing the versine 

vers 20 = 1 - cos 20 = 2 sin 2 = l 2 /2 


where Q = l/d 2 , one can use tables of sin a 
(and cos a), « = 0(0.01°)90°, as follows: 
(1) for 0<2G^90°, let * = 20, look up cos 
x and take its complement to 1; (2) for 90° < 
29<180°, let y = 29-90°, so that vers 20 = 
1 + sin y, look up sin y and add to 1. 

(1295) Calibration sights for X-ray powder 
camera. G. Donnay and J. G. Smith. 
Am. Mineralogist, 44, 196-199 (1959). 

When measuring a powder film it is neces- 
sary to find the midpoints of incident and 
emerging X-ray beams and also to determine 
the film shrinkage. Calibration sights are at- 
tached to a flat ring l /& inch thick, made of 
aluminum alloy and of such a size as to fit 
against the shoulder on the inside of the 
Buerger-type powder camera. The sights 
have a twofold purpose. They permit deter- 
mination of film shrinkage by direct meas- 
urement, provided the film has been devel- 
oped with care to ensure uniform shrinkage. 
They permit centering the film on the meas- 
uring device so that the midpoint of the 
X-ray diffraction rings comes at a convenient 
reading, say 100.00 mm. This procedure elim- 
inates the need for finding the midpoint 
by reading ring positions on both sides of the 
punched hole in the film. It simplifies the 
arithmetic of subtracting the midpoint read- 
ing from the reading of each ring. 

(1296) Bone doses from strontium-90. W. F. 
Libby. Proc. Natl. Acad. Sci. U. S., 
45,245-249 (1959). 

The calculation of the radiation doses re- 
ceived by bone marrow from Sr 90 is made in 
a new way utilizing a methodology developed 
by the author and his students earlier for 
absolute beta radiation counting. In principle 
the calculation is made on the assumption 
that the intensity decreases exponentially with 
the thickness of absorber. Under conditions 
thought to be applicable to the bone dose 
problem, this assumption should be satisfac- 
tory. A general equation is developed for the 
dose to layers of marrow from Sr 00 deposited 
in bone layers bounding the marrow on 
either side in a sandwich type of arrangement. 
The closed expression for the dose rate is 
compared with earlier and more complicated 

(1297) Pyrite stability relations in the Fe-S 
system. G. Kullerud and H. S. Yoder. 
Econ. Geol, 54, 533-572 (1959). 

The reaction pyrite =± pyrrhotite + liquid 
(or gas) was investigated up to 5000 bars by 
means of a new technique. The univariant 
equilibrium curve for the reaction originates 
at an invariant point 743° C and about 10 
bars and passes through the points 748° C, 
335 bars; 755°, 1000 bars; 770°, 2000 bars; 
and 810°, 5000 bars. 

The relations of pyrite in the Fe-S system 
are deduced from thermodynamic principles 
and available data. The limitations of the 
various experimental techniques for sulfide 
systems are analyzed in the light of the pres- 
sure-temperature diagram for the Fe-S sys- 
tem. The apparently large effect of iron in 
the vapor or gas of the system is described. 

The occurrence of primary pyrite in some 
gabbros, amphibolites, and granites and not 
in basalts and rhyolites is accounted for by 
comparing the upper stability curves of pyrite 
at various partial pressures of sulfur with the 
beginning-of-melting curves of rocks of ba- 
saltic and granitic compositions under hy- 
drous conditions. When the partial pressure 
of sulfur is less than about 10 bars small 
changes in its value have a great effect on the 
stability of pyrite. The relations of pyrite in 
the Fe-S system indicate that massive pyrite 
bodies could not have crystallized directly 
from a liquid of pyrite composition. 

(1298) Magnetic properties of solid solutions of 
Fe 3 4 and FeAl 2 4 . S. J. Pickart and 
A. C. Turnock. /. Phys. Chem. Solids, 
10,242-243 (1959). 

The spinel-type solid solution series from 
magnetite (Fe 3 4 ) to hercynite (FeAl 2 4 ) 
has been synthesized hydrothermally. The 
saturation magnetizations of various compo- 
sitions were measured in fields up to 6300 
oersteds from 77° K to the Curie point. The 
plot of the moments indicates that the struc- 
ture changes continuously from "inverse" 
Fe 3 4 to "normal" FeAl 2 4 . The magneti- 
zation curves are of Neel's P type, with a 
transition to the N type; the compensation 
point is at about r=1.2 (FeFe 2 _jAlj0 4 ). 


(1299) The crystal structure of hexagonal 
CaAl 9 Si 2 O g . Y. Takeuchi and G. Don- 
nay. 'Acta Cry st., 12, 465-470 (1959). 

The structure of hexagonal CaAl 2 Si 2 8 has 
been determined by two-dimensional Fourier 
methods. Double sheets of composition Al 2 - 
Si 2 O s are made up of linked oxygen tetra- 
hedra about (Si,Al) atoms. They are similar 
to the sheets found in hexagonal BaAl 2 Si 2 8 , 
but are considerably deformed from hexago- 
nal to trigonal symmetry, in such a way as to 
contract the cationic niches between the sheets. 
A comparison of the two structures illustrates 
the nonrigidity of the anionic layers. 

Diffuse streaks on the Weissenberg photo- 
graphs indicate mistakes in the stacking of the 
sheets. Such mistakes lead to the formation 
of trigonal prisms of oxygen atoms around 
calcium atoms, whereas in the ordered struc- 
ture calcium is coordinated octahedrally. The 
Ca-O distance is found to be 2.39 A in either 
coordination. The (Si,Al)-0 distances are 
1.66 and 1.71, both ±0.03 A. The residual, 
i?=13.8 per cent, refers to the 151 permitted 
nonequivalent reflections in the Cu Ka sphere, 
49 of which are structurally absent. 

(1300) Geochemistry of organic substances. 
P. H. Abelson. In Researches in Geo- 
chemistry (ed. P. H. Abelson), John 
Wiley & Sons, New York, pp. 79-103, 

This paper is a discussion of kinds of sub- 
stances produced by living matter and the fate 
of these substances in geologic environments. 

(1301) Mineral assemblages of the Green River 
formation. C. Milton and H. P. Eug- 
ster. In Researches in Geochemistry 
(ed. P. H. Abelson), John Wiley & Sons, 
New York, pp. 118-150, 1959. 

Origin and nature of the Green River for- 
mation are discussed in an introductory sec- 
tion, followed by a table of all authigenic min- 
erals found to date in the Green River beds. 
This table, assigning minerals to the three 
main basins, contains several new minerals 
and others found elsewhere only in differing 
geologic environments. The genetic signifi- 
cance of mineral assemblages found is dis- 
cussed, with chief emphasis on the alkali car- 
bonates, trona, nahcolite, shortite, eitelite, and 
others. Pqoo~T phase diagrams, constructed 
from published data, are used in evaluating 

environmental differences between the Utah, 
Wyoming, and Colorado basins. Other min- 
erals discussed include analcite, searlesite, 
reedmergnerite, acmite, riebeckite, sepiolite, 
loughlinite, elpidite, labuntsovite, leucosphe- 
nite, pyrite, and bradleyite. 

(1302) Tritium in hydrology and meteorology. 
W. F. Libby. In Researches in Geo- 
chemistry (ed. P. H. Abelson), John 
Wiley & Sons, New York, pp. 151-168, 

The general principles of the use of cosmic- 
ray and nuclear bomb tritium in the study of 
hydrology and atmospheric geophysics are 
presented and discussed as they were devel- 
oped on the basis of the data obtained in the 
Chicago laboratories. The conclusion that 
tritium hydrology is a new tool of real po- 
tentiality seems to be justified, and possibili- 
ties of using synthetic labeling for large 
bodies of water are pointed out and discussed. 

(1303) Geochronology. G. R. Tilton and G. L. 
Davis. In Researches in Geochemistry 
(ed. P. H. Abelson), John Wiley & 
Sons, New York, pp. 190-216, 1959. 

Mass spectrometers and isotopic dilution 
methods for the determination of the radio- 
active parents and daughter isotopes for age 
determination have been used for ten years. 
The application to geological problems of 
the geochronological information so far ob- 
tained is reviewed with respect to reliability 
of the measurements, reliably dated localities 
throughout the world, study of orogenic belts, 
metamorphic rocks, the fossil time scale, and 
the interpretation of discordant ages. 

(1304) Sulfide systems as geological thermom- 
eters. G. Kullerud. In Researches in 
Geochemistry (ed. P. H. Abelson), John 
Wiley & Sons, New York, pp. 301-335, 

This paper discusses the stability of such 
minerals as pyrite and covellite as well as the 
stability relations of some of the mineral as- 
semblages in the copper-iron-sulfur and sul- 
fur-iron-oxygen systems. Solid solutions 
among minerals such as pyrrhotite-pyrite, 
pyrrhotite-sphalerite, and pyrite-sphalerite are 
considered. Examples are given of applica- 
tions of these relations to sulfides in rocks and 
ore bodies. 


(1305) Diffraction effects of short-range order- 
ing in layered sequences. F. Chayes. 
In Researches in Geochemistry (ed. 
P. H. Abelson), John Wiley & Sons, 
New York, pp. 359-376, 1959. 

Numerical relations between run frequen- 
cies, run lengths, and short-range ordering 
in simple arrays containing items of two kinds 
are outlined. A method of generating se- 
quences characterized by any desired level of 
short-range ordering by controlling run fre- 
quencies and run lengths is presented. The 
construction of two-dimensional diffraction 
masks exhibiting different levels of short- 
range ordering and preliminary experimental 
study of diffraction by such masks are de- 
scribed. The possibility of substituting high- 
speed digital calculation for optical experi- 
mentation is briefly reviewed. 

(1306) Hydrothermal investigations of amphi- 
boles. F. R. Boyd. In Researches in 
Geochemistry (ed. P. H. Abelson), John 
Wiley & Sons, New York, pp. 377-396, 

Hydrothermal investigations of amphiboles 
have been undertaken to provide quantita- 
tive data on the upper boundary of the am- 
phibolite facies. Chemical studies of natural 
calciferous amphiboles have shown that trem- 
olite-ferrotremolite, Ca 2 (Mg,Fe) 5 Si 8 02 2 - 

(OH) 2 , and pargasite-ferropargasite, NaCa 2 - 
(Mg,Fe) 4 AlAl 2 Si 6 22 (OH) 2 , are the most 
important end member series of this group. 
Experimentally determined phase diagrams 
for tremolite and pargasite are presented. 
Tremolite breaks down to enstatite + diopside 
+ quartz + vapor at 835° C at 1000 bars; at 
the same pressure pargasite is stable up to 
1040° C. These data help explain the fact 
that a hornblende is often found in granulites 
and quartz diorites along with hypersthene + 
augite + quartz, which are the breakdown 
products of an intermediate member of the 
tremolite-ferrotremolite series. Experimental 
data show that pure magnesian pargasite is 
not stable in the presence of quartz, in agree- 
ment with natural occurrences of this mineral. 

Experimental data obtained by various 
authors for the anthophyllite-cummingtonite 
group are reviewed. No stability field has yet 
been established for these amphiboles, though 
magnesian anthophyllite and anthophyllites 
and cummingtonites of intermediate Fe/Mg 
ratio have been synthesized. 

The stability field of the alkali amphibole 
magnesian riebeckite, ° Na 2 Mg 3 Fe 2 "'Si 8 22 - 
(OH) 2 , has been determined by W. G. 
Ernst. Ernst has also synthesized ferrorie- 
beckite and glaucophane. Preliminary data 
indicate that there is no major difference in 
the Ph 2 o~T ranges of stability of calciferous 
and alkali amphiboles. 

(1307) Reduction and oxidation in metamor- 
phism. H. P. Eugster. In Researches 
in Geochemistry (ed. P. H. Abelson), 
John Wiley & Sons, New York, pp. 397- 
426, 1959. 

This paper analyzes the significance of re- 
dox reactions in geologic environments, par- 
ticularly during metamorphism. A brief re- 
view of the systems Fe-O, Fe-Si-O, Fe-Si- 
O-H presents the data so far available. The 
stability of hydrous iron silicates is discussed 
by using new data on the stability of annite, 
KFe 3 AlSi 3 Oio(OH) 2 , an iron biotite, as an 
example. An experimental method is briefly 
reviewed for controlling the partial pressure 
of oxygen at high vapor pressures and temper- 
atures. Data on the stability of other hydrous 
iron silicates are briefly presented. Geologic 
applications discussed include sections on 
metamorphism of Precambrian iron forma- 
tion, on redox reactions and mineral facies, 
and on reduction and oxidation during meta- 

(1308) Equations of state and polymorphism 
at high pressures. S. P. Clark, Jr. In 
Researches in Geochemistry (ed. P. H. 
Abelson), John Wiley & Sons, New 
York, pp. 495-511, 1959. 

Experimental evidence for pressure-induced 
phase changes in silicates is reviewed, and it 
is concluded that some of the most drastic 
effects of pressure may involve heterogeneous 
equilibria in multicomponent systems. Elec- 
tronic breakdown at very high pressure is 
discussed, and two equations of state are pre- 
sented. One, based on finite-strain theory, is 
valid at low pressures where experimental 
data are available. The quantum-statistical 
theory of Thomas and Fermi is valid only at 
extremely high pressure. Comparison of the 
two theories in the case of the alkali metals 
suggests that the Thomas-Fermi theory does 
not give reliable results at pressures less than 
about 10 8 bars. Geophysical implications of 


some of these considerations are briefly dis- 

(1309) Nepheline solid solutions. G. Donnay, 
J. F. Schairer, and J. D. H. Donnay. 
Mineralog. Mag., 32, 93-109 (1959). 

Published chemical analyses demonstrate 
that the nepheline formula should be written 

i^zNaj / Ca 2 n 8-(a; + y + Z)Altf + j / + 2sOl 1 6-(tf; + j/ + 22)U32, 

where D stands for vacant sites. X-ray data are 
presented for the nepheline phase in four 
binary systems: Ne-CaAl 2 4 , Ne-An, Ne- 
Ab, Ne-Kp. Only in two of these systems do 
the cell dimensions change with composition. 
In the first one, the cell volume V increases 
linearly with increasing calcium content; in 
the last one, two singularities in the curve of 
V against x divide the phase Na 8 _a;K^Al 8 Si 8 - 
32 into three subphases: subpotassic (0< 
at<0.25), mediopotassic (0.25<at<2.00), 
and perpotassic (2.00<*<4.73). Both highl- 
and low-temperature forms are found in the 
subpotassic range only. Twenty-eight natural 
nephelines, for which chemical analyses and 
X-ray data are available in the literature, 
show that only the potassium content affects 
cell dimensions. Although all analyzed nat- 

ural nephelines fall outside the subpotassic 
range, the re-examination of a Monte Somma 
specimen studied by Bannister reveals a few 
euhedral crystals of subpotassic nepheline in 
a mediopotassic phase. 

(1310) Paleobiochemistry and organic geo- 
chemistry. P. H. Abelson. Fortschr. 
Chem. org. Naturstoffe, 17, 379-403, 

This paper discusses some recent discoveries 
in comparative biochemistry and their rela- 
tion to paleobiochemistry, and the degradation 
of organic matter in sediments by biological 
and chemical mechanisms. 

(1314) Tables de groupes spatiaux magnetiques. 
J. D. H. Donnay and G. Donnay. 
Compt. rend. acad. sci. Paris, 248, 3317- 
3319 (1959). 

The 38 Shubnikov space groups that stem 
from the point group 4/m are tabulated as 
an example of the mode of presentation. In 
each space group the table gives the axial 
vector (magnetic moment) [uvw] associated 
with each site (xyz) of the general position. 


Abelson, P. H. Geochemistry of organic sub- 
stances. In Researches in Geochemistry (ed. 
P. H. Abelson), John Wiley & Sons, New 
York, pp. 79-103, 1959. 

Abelson, P. H. Paleobiochemistry and organic 
geochemistry. Fortschr. Chem. org. Natur- 
stoffe, 17, 379-403 (1959). 

Aldrich, L. T, G. W. Wetherill, G. L. Davis, and 
G. R. Tilton. Radioactive ages of micas 
from granitic rocks by Rb-Sr and K-A meth- 
ods. Trans. Am. Geophys. Union, 39, 1 124 — 
1134 (1958). 

Bowen, N. L., See Tuttle, O. F. 

Boyd, F. R. Hydrothermal investigations of 
amphiboles. In Researches in Geochemis- 
try (ed. P. H. Abelson), John Wiley & Sons, 
New York, pp. 377-396, 1959. 

Chayes, F. Diffraction effects of short-range 
ordering in layered sequences. In Re- 
searches in Geochemistry (ed. P. H. Abel- 
son), John Wiley & Sons, New York, pp. 
359-376, 1959. 

Clark, S. P., Jr. Equations of state and poly- 
morphism at high pressures. In Researches 
in Geochemistry (ed. P. H. Abelson), John 
Wiley & Sons, New York, pp. 495-511, 1959. 

Corliss, L. M. See Donnay, G. 

Davis, G. L. See Aldrich, L. T.; Tilton G. R. 

Donnay, G. Crystal data on chlorophyll a. 
Arch. Biochem. Biophys., 80, 80-85 (1959). 

Donnay, G., L. M. Corliss, J. D. H. Donnay, N. 
Elliott, and J. M. Hastings. Symmetry of 
magnetic structures: Magnetic structure of 
chalcopyrite. Phys. Rev., 112, 1917-1923 

Donnay, G., J. F. Schairer, and J. D. H. Don- 
nay. Nepheline solid solutions. Mineralog. 
Mag., 32, 93-109 (1959). 

Donnay, G., and J. G. Smith. Calibration sights 
for X-ray powder camera. Am. Mineral- 
ogist, 44, 196-199 (1959). 

Donnay, G. See also Donnay, J. D. H.; Kul- 
lerud, G.; Takeuchi, Y. 

Donnay, J. D. H., and G. Donnay. Sine table 
for indexing powder patterns. Am. Min- 
eralogist, 44, 177-179 (1959). 

Donnay, J. D. H., and G. Donnay. Tables de 
groupes spatiaux magnetiques. Compt. rend, 
acad. sci. Paris, 248, 3317-3319 (1959). 

Donnay, J. D. H. See also Donnay, G. 

Elliott, N. See Donnay, G. 

Eugster, H. P. Reduction and oxidation in met- 
amorphism. In Researches in Geochemistry 
(ed. P. H. Abelson), John Wiley & Sons, 
New York, pp. 397-426, 1959. 

Eugster, H. P. See also Milton, C. 



Hastings, J. M. See Donnay, G. 

Hopson, C. A. See Tilton, G. R. 

Kullerud, G. Sulfide systems as geological ther- 
mometers. In Researches in Geochemistry 
(ed. P. H. Abelson), John Wiley & Sons, 
New York, pp. 301-335, 1959. 

Kullerud, G., and G. Donnay. Natural and syn- 
thetic ferroselite: A roentgenographic mime- 
sis of rammelsbergite. Geochim. et Cos- 
mochim. Acta, 15, 73-79 (1958). 

Kullerud, G., and H. S. Yoder. Pyrite stability 
relations in the Fe-S system. Econ. Geol., 
54, 533-572 (1959). 

Libby, W. F. Bone doses from strontium-90. 
Proc. Natl. Acad. Sa. U. S., 45, 245-249 

Libby, W. F. Tritium in hydrology and me- 
teorology. In Researches in Geochemistry 
(ed. P. H. Abelson), John Wiley & Sons, 
New York, pp. 151-168, 1959. 

Milton, C., and H. P. Eugster. Mineral as- 
semblages of the Green River formation. In 
Researches in Geochemistry (ed. P. H. Abel- 
son), John Wiley & Sons, New York, pp. 
118-150, 1959. 

Pickart, S. J., and A. C. Turnock. Magnetic 
properties of solid solutions of Fe 3 4 and 
FeAh0 4 . /. Phys. Chem. Solids, 10, 242- 
243 (1959). 

Schairer, J. F. Phase equilibria with particular 
reference to silicate systems. Chapter 5 in 
The Technique of Physico-Chemical Meas- 
urements at High Temperatures (ed. J. O'M. 
Bockris, J. D. Mackenzie, and J. L. White), 

Butterworths Scientific Publications, Lon- 
don, 1959. 

Schairer, J. F. See also Donnay, G. 

Smith, J. G. See Donnay, G. 

Smith, J. R. The optical properties of heated 
plagioclases. Am. Mineralogist, 43, 1 179— 
1194 (1958). 

Takeuchi, Y., and G. Donnay. The crystal struc- 
ture of hexagonal CaAl Si 2 8 . Acta Cryst., 
12, 465-470 (1959). 

Tilton, G. R. Isotopic composition of lead from 
tektites. Geochim. et Cosmochim. Acta, 14, 
323-330 (1958). 

Tilton, G. R., and G. L. Davis. Geochronology. 
In Researches in Geochemistry (ed. P. H. 
Abelson), John Wiley & Sons, New York, 
pp. 190-216, 1959. 

Tilton, G. R., G. W. Wetherill, G. L. Davis, and 
C. A. Hopson. Ages of minerals from the 
Baltimore gneiss near Baltimore, Maryland. 
Bull. Geol. Soc. Am., 69, 1469-1474 (1958). 

Tilton, G. R. See also Aldrich, L. T. 

Turnock, A. C. See Pickart, S. J. 

Tuttle, O. F., and N. L. Bowen. Origin of 
granite in the light of experimental studies 
in the system NaAlSi 3 8 -KAlSi 3 8 -Si0 2 - 
H 2 0. Geol. Soc. Am. Mem. 74, xi + 153 pp., 
67" figs., 6 pi. (1958). 

Wetherill, G. W. See Aldrich, L. T.; Tilton, 
G. R. 

Yoder, H. S., Jr. Experimental studies on micas: 
A synthesis. Proc. Sixth Natl. Con], on 
Clays and Clay Minerals, 42-60 (1959). 

Yoder, H. S., Jr. See also Kullerud, G. 


Agrell, S. O., and J. M. Langley. The dolerite 
plug at Tievebullaigh, County Antrim. 
Proc. Roy. Irish Acad., B, 59, 93-127 (1958). 

Alderman, A. R. Eclogites from the neighbour- 
hood of Glenelg, Inverness-shire. Quart. }. 
Geol. Soc. London, 92, 488-530 (1936). 

Atlas, L. The polymorphism of MgSiO s and 
solid state equilibria in the system MgSiO s - 
CaMgSi 2 6 . /. Geol, 60, 125-147 (1952). 

Atlas, L. M., and W. K. Sumida. Solidus, sub- 
solidus, and subdissociation phase equilibria 
in the system Fe-Al-O. /. Am. Ceram. Soc, 
41, 150-160 (1958). 

Belov, N. V., N. N. Neronova, and T. S. Smir- 
nova. The 1651 Shubnikov groups. Trudy 
Inst. Krist., A\ad. Nau\ SSSR, 11, 33-67 
(1955) (in Russian). 

Birch, F. Physics of the crust. Geol. Soc. Am. 
Spec. Paper 62, 101-118 (1955). 

Birch, F., and P. LeComte. Plane of albite 
composition to 1000° and 26,000 bars. Ab- 
stract of paper presented at meetings of the 
American Association for the Advancement 

of Science, Washington, D. C, December 

Bornemann, K. Das Schmelzdiagramm der 

Nickel-Schwefelverbindungen. Metallurgie, 

5, 13-19 (1908). 
Bowen, N. L. The behavior of inclusions in 

igneous magma. /. Geol., 30, 551 (1922). 
Bowen, N. L. The Evolution of the Igneous 

Rocl{s. Princeton University Press, 1928 (re- 
print by Dover Publications, 1956). 
Bowen, N. L., and R. B. Ellestad. Leucite and 

pseudoleucite. Am. Mineralogist, 22, 567- 

571 (1937). 
Bowen, N. L., and J. W. Greig. The system 

Al 9 3 -Si0 2 . /. Am. Chem. Soc, 7, 238-254 

Bowen, N. L., and O. F. Tuttle. The system 

NaAlSi 3 8 -KAlSi 3 8 -H 2 0. /. Geol, 58, 

489-511 (1950). 
Bown, M. G., and P. Gay. Observation on pi- 

geonite. Acta Cryst., 10, 440 (1957). 
Bridgman, P. W. Polymorphism, principally of 

the elements, up to 50,000 kg/cm 2 . Phys. 

Rev., 48, 893-906 (1935). 


Bridgman, P. W. The resistance of 72 elements, 
alloys, and compounds to 100,000 kg/cm 2 . 
Proc. Am. Acad. Arts Sci., 81, 165-251 

Buerger, M. J., and W. Parrish. The unit cell 
and space group of tourmaline (an example 
of the inspective equi-inclination treatment 
of trigonal crystals). Am. Mineralogist, 22, 
1139-1150 (1937). 

Bugge, J. S. W. Geological and petrographical 
investigations in the Kongsberg-Bamble for- 
mation. Norg. Geol. Undersokflse, no. 160 

Buryanova, E. Z., and A. I. Komkov. A min- 
eral — ferroselite. Do\lady A\ad. Nau\ 
SSSR, 105, 812-813 (1955). 

Carslaw, H. S., and J. C. Jaeger. Conduction of 
Heat in Solids. Oxford University Press, 
386 pp., 1947. 

Clark, L. A. Phase Relations in the Fe-As-S 
System. Unpublished Ph.D. dissertation, 
McGill University, 1959. 

Clark, S. P., E. C. Robertson, and F. Birch. 
Experimental determination of kyanite-silli- 
manite equilibrium relations at high temper- 
atures and pressures. Am. }. Sci., 255, 628- 
640 (1957). 

Donnay, G. Hexagonal CaAl 2 Si 2 8 . Acta 
Cryst.,5, 153 (1952). 

Donnay, G., N. V. Belov, N. N. Neronova, and 
T. S. Smirnova. On Shubnikov groups. 
Kristallografiya, 3, 635-636 (1958) (in Rus- 

Donnay, G., J. D. H. Donnay, and G. Kullerud. 
Crystal and twin structure of digenite, Cu 9 - 
S 5 . Am. Mineralogist, 43, 230-242 (1958). 

Donnay, J. D. H., and G. Donnay. Tables de 
groupes spatiaux magnetiques. Compt. rend. 
acad. sci. Paris, 248, 3317-3319 (1959). 

Emerson, B. K. Geology of Massachusetts and 
Rhode Island. U. S. Geol. Survey Bull. 597, 
188-193 (1917). 

Eskola, P. On the relations between the chemi- 
cal and mineralogical composition in the 
metamorphic rocks of the Orijarvi region. 
Bull. comm. geol. Finlande, 44 (1915). 

Eskola, P. Orijarvi re-interpreted. Bull. comm. 
geol. Finlande, 150, 93-102 (1950). 

Eskola, P. On the granulites of Lapland. Am. 
J. Sci., Bo wen Vol., 133-171 (1952). 

Eugster, H. P. Heterogeneous reactions involv- 
ing oxidation and reduction at high pressures 
and temperatures. /. Chem. Phys., 26, 1760- 
1761 (1957). 

Everenden, }. F., G. H. Curtis, R. Kistler, and 
J. D. Obradovich. Argon diffusion in K 
feldspar, biotite, glauconite, and illite. Ab- 
stract of paper presented at American Geo- 
physical Union Meetings, May 1959. 

Fermor, L. L. Preliminary note on garnet as a 
geological barometer and on an infra-plu- 

tonic zone in the earth's crust. Records Geol. 
Survey India, 43, 41-47 (1913). 

Fermor, L. L. Discussion of Tilley, C. E.: 
Contact-metamorphism in the Comrie area 
of the Perthshire Highlands. Quart, f. Geol. 
Soc. London, 80, 70-71 (1924). 

Fischer, G. Uber das Grundgebirge der Bayer- 
ischen Ostmark. Die Gneise nordlich des 
Bayerischen Pfahles. Jahrb. preuss. geol., 
Landesanstalt Berlin, 59 (1938). 

Fleischer, M. The relation between chemical 
composition and physical properties in the 
garnet group. Am. Mineralogist, 22, 751- 
759 (1937). 

Ford, W. E. A study of the relations existing 
between the chemical, optical, and other 
physical properties of members of the garnet 
group. Am. J. Sci., 40, 33-49 (1915). 

Forsman, J. P., and J. M. Hunt. Insoluble or- 
ganic matter (kerogen) in sedimentary 
rocks. Geochim. et Cosmochim. Acta, 15, 
170-182 (1958). 

Foster, W. R. High-temperature X-ray diffrac- 
tion study of the polymorphism of MgSiO s . 
/. Am. Ceram. Soc, 34, 255-259 (1951). 

Friedman, G. M. The spinel-silica reaction suc- 
cession: A study of incompatible mineral 
phases. /. Geol, 62, 366-374 (1954). 

Friedman, G. M. The origin of spinel-emery 
deposits with particular reference to those 
of the Cortlandt complex, New York. N. Y. 
State Museum. Bull. 351 (1956). 

Friedrich, K., and F. Benningson. Nickel und 
Arsen. Metallurgie, 4, 200-216 (1907). 

Frueh, A. J., Jr. Disorder in the mineral born- 
ite, Cu r ,FeS 4 . Am. Mineralogist, 35, 185— 
192 (1950). 

Gerling, E. K., and I. M. Morozova. Deter- 
mination of the activation energy of argon 
isolation from micas. Geochemia, 304-311 

Griggs, D. T., and G. C. Kennedy. A simple ap- 
paratus for high pressures and temperatures. 
Am. J. Sci., 254, 722-735 (1956). 

Harker, A. Metamorphism. 2d ed., London, 
Methuen and Co., 1939. 

Heyding, R. D., and L. D. Calvert. Arsenides 
of the transition metals. II. The nickel ar- 
senides. Can. J. Chem., 35, 1205-1215 

Hodge, J. E. Dehydrated foods: Chemistry of 
browning reactions in model systems. Agr. 
Food Chem., 1, 928-943 (1953). 

Hoffmann, A., and W. A. Fischer. Bildung des 
Spinells Fe0Al 2 3 und seiner Mischkristalle 
mit Fe 3 4 und A1 2 3 . Z. physi\. Chem., 
N.F. 7, 80-90 (1956). 

Iiyama, T. Existence of indialite solid solutions. 
Proc. Japan Acad., 31, 166-168 (1955). 

Ito, T. X-ray Studies in Polymorphism. Ma- 
ruzen, Tokyo, 1950. 



de Jong, W. F., and H. W. V. Willems. Exis- 
tenz und Struktur der Disulfide NiS 2 und 
CoSo. Z. anorg. Chan., 160, 185-189 

Keith, M. L., and J. F. Schairer. The stability 
field of sapphirine in the system MgO- 
Al 2 3 -Si0 2 . /. Geol, 60, 181-186 (1952). 

King, P. B. The Tectonics of Middle North 
America. Princeton University Press, 203 
pp., 1951. 

Kracek, F. C. Phase relations in the system sul- 
fur-silver and the transitions in silver sul- 
fide. Trans. Am. Geophys. Union, 27 , 364— 
374 (1946). 

Kullerud, G. The FeS-ZnS system: A geological 
thermometer. Nors^. Geol. Tidss\r., 32, 
61-147 (1953). 

Kullerud, G., and E. H. Roseboom, Jr. Cu 9 S 5 - 
Cu 5 FeS 4 system. Bull. Geol. Soc. Am., 69, 
1602 (1958). 

La Forge, L. Geology of the Boston area, Mas- 
sachusetts. U. S. Geol. Survey Bull. 839, 
36 (1932). 

Libby, W. F. Radioactive strontium fallout. 
Proc. Natl. Acad. Sci. U. S., 42, 365-390 

Libby, W. F. Current research findings on ra- 
dioactive fallout. Proc. Natl. Acad. Sci. 
U. S., 42,945-962 (\956b). 

Libby, W. F. Radioactive fallout. Proc. Natl. 
Acad. Sci. U. S., 43, 758-775 (1957). 

Libby, W. F. Radioactive fallout. Bull. Swiss 
Acad. Med. Sci., 14, 309 (1958), or Proc. 
Natl. Acad. Sci. U. S., 44, 800 (1958). 

Loeb, A. L. A binary algebra describing crystal 
structures with closely packed anions. 
Acta Cryst., 11, 469-476 (1958). 

Lundquist, D., and A. Westgren. On the crystal 
structure of bornite, Cu 5 FeS 4 . Ar\iv Kemi 
Mineral. Geol, B, 12, no. 23, 6 (1936). 

Machta, L. Nature of radioactive fallout and 
its effects on man. Congressional Hearings 
before the Joint Committee on Atomic 
Energy, Part I, p. 141, June 1957. 

Machta, L. Meteorological interpretation of 
strontium-90 fallout. Presented at public 
meeting sponsored by Washington Chapter 
of Federation of American Scientists, May 1, 
1958, and published in HASL-42 October 

Machta, L. Transport in the stratosphere and 
through the tropopause. Advances in Geo- 
phys., 6, 273 (1959). 

MacKenzie, W. S. The crystalline modifications 
of NaAlSi 3 O s . Am. J. Sci., 255, 481-516 

Martell, E. A. Atmospheric aspects of stron- 
tium-90 fallout. Science, 129, 1197-1206 

Miyashiro, A. Cordierite-indialite relations. 
Am. J. Sci., 255, 43-62 (1957). 

Miyashiro, A., and T. Iiyama. A preliminary 
note on a new mineral, indialite, poly- 
morphic with cordierite. Proc. Japan Acad., 
30, 746-751 (1954). 

Montgomery, A. Precambrian geology of the 
Picuris Range, north central New Mexico. 
New Mexico Bur. Mines and Mineral Re- 
sources Bull. 30, 89 pp. (1953). 

Mora, P. T., and J. W. Wood. Synthetic poly- 
saccharides. I. Polycondensation of glu- 
cose. /. Am. Chem. Soc, 80, 685-692 (1958). 

Morey, G. W. The system water-nepheline-al- 
bite: A theoretical discussion. Am. J. Sci., 
255, 461-480 (1957). 

Morimoto, N. The existence of monoclinic py- 
roxenes with the space group C 2h 5 -P2 1 /c. 
Proc. Japan Acad., 32, 750-752 (1956). 

Morimoto, N., and T. Ito. Pseudo-twin of augite 
and pigeonite. Bull. Geol. Soc. Am., 69, 
1616 (1958). 

Nelson, B. W., and R. Roy. Synthesis of the 
chlorites and their structural and chemical 
constitution. Am. Mineralogist, 43, 707-725 

Newhouse, W. H., and A. F. Hagner. Cordierite 
deposits of the Laramie Range, Albany 
County, Wyoming. Wyoming Geol. Survey 
Bull. 41 (1949). 

Nockolds, S. R. Average chemical composition 
of some igneous rocks. Bull. Geol. Soc. 
Am., 65, 1007-1032 (1954). 

Okazaki, A., and K. Hirakawa. The X-ray study 
of FeSe^,. II. Busseiron Ken\yu, 90, 59-60 

Peacock, M. A. On maucherite (nickelspeiss, 
placodine, temiskamite). Mineralog. Mag., 
25,557-572 (1940). 

Poldervaart, A., and H. H. Hess. Pyroxenes in 
the crystallization of basalt magma. /. 
Geol, 59, m-4%9 (1951). 

Ramdohr, P. Die Erzmineralien und ihre 
Verivachsungen. Berlin, Akademie-Verlag, 

Rankin, G. A., and H. E. Merwin. The ternary 
system MgO-Al,0 3 -Si0 2 . Am. J. Sci., 45, 
301-325 (1918)." 

Raudsepp, Kh. T. Concerning a new method of 
investigation of the chemical composition of 
mineral fuels and concerning the chemical 
structure of Estonian oil shale — kukersite. 
Izvest. A\ad. Nau\ SSSR, Otdel. Te\h. 
Nau\, no. 3 (1954). 

Read, H. H. On corundum-spinel xenoliths in 
the gabbro of Haddo House, Aberdeenshire. 
Geol. Mag., 68, 446-453 (1931). 

Reamer, H. H., B. H. Sage, and W. N. Lacey. 
Volumetric behavior of hydrogen sulfide. 
Ind. Eng. Chem., 42, 140-143 (1950). 

Robertson, E. C, F. Birch, and G. J. F. Mac- 
Donald. Experimental determination of 


jadeite stability relations to 25,000 bars. 
Am. J. Set., 255, 115-137 (1957). 

Schairer, J. F. The system K 2 0-MgO-Al 2 3 - 
SiO,: I. Results of quenching experiments 
on four joins in the tetrahedron cordierite- 
forsterite-leucite-silica and on the join cor- 
dierite-mullite-potash feldspar. /. Am. 
Ceram. Soc, 37, 501-533 (1954). 

Scheumann, K. H. Das Problem der Cordierit- 
Gneise im Sachsischen Granulitgebirge. 
Mineral, u. petrog. Mitt., N. F. 53, 147-154 

Selleck, F. T., L. T. Carmichael, and B. H. Sage. 
Phase behavior in the hydrogen sul fide-water 
system. Ind. Eng. Chem., 44, 2219-2226 

Shelton, Frank. High-altitude sampling pro- 
gram. Public Hearings on Fallout from 
Nuclear Weapons Tests. Testimony before 
the Special Committee on Atomic Energy, 
Special Subcommittee on Radiation, Joint 
Committee on Atomic Energy, May 5-8, 
1959. U. S. Government Printing Office (in 

Shibata, H. Graphic intergrovvths of cordierite 
and quartz in pegmatites from Sasago and 
Dosi, Province of Kai, Japan. Japan J. Geol. 
and Geogr., 13, 205-227 (1936). 

Skinner, B. J. Physical properties of end mem- 
bers of the garnet group. Am. Mineralogist, 
41, 428-436 (1956). 

Smith, J. V. Crystal structure of protoenstatite. 
Bull. Geol. Soc. Am., 69, 1645 (1958). 

Stillwell, F. L., and A. B. Edwards. Petrology 
of the Gumeracha talc deposits. Geol. Soc. 
S. Australia Bull. 26, 31-49 (1951). 

Suess, F. E. Der Granulitzug von Borry in 
Mahren. Jahrb. geol. Reich sanstalt, Wien, 
50, 615-648 (1900). 

Sundius, N. The classification of the horn- 
blendes and the solid solution relations in 
the amphibole group. Sveriges Geol. Un- 
der so\n. Arsbo\, 40, no. 4, 1-36 (1946). 

Takeuchi, Y. A detailed investigation of the 
structure of hexagonal BaAl 2 Si 2 O s with ref- 
erence to its a-3 inversion. Mineral, f. 
(Sapporo), 2, 311-332 (1958). 

Thompson, J. B. Mineral assemblages in pelitic 
schists. Am. Mineralogist, 42, 842-858 

Tilley, C. E. Contact metamorphism in the 
Comrie area, Perthshire Highlands. Quart. 
J. Geol. Soc. London, 80, 22-71 (1924). 

Tilley, C. E. Problems of alkali rock genesis. 
Quart. J. Geol. Soc. London, 113, 323-360 

Travis, R. B. Note on large cordierite porphy- 
roblasts, Fremont County, Colorado. Am. 
Mineralogist, 41, 796-799 (1956). 

Tunell, G., and C. E. Adams. Bornite of Cam 
Brea Mine, Illogan, Cornwall, Toronto. 
Am. Mineralogist, 34, 824-829 (1949). 

Tuttle, O. F., and N. L. Bowen. Origin of 
granite in the light of experimental studies 
in the system NaAlSi 3 8 -KAlSi.,O s -SiOo- 
H 2 0. Geol. Soc. Am. Mem. 74 (1958). 

Warren, B. E., and J. Biscoe. The crystal struc- 
ture of the monoclinic pyroxenes. Z. Krist., 
80, 39M01 (1931). 

Warren, B. E., and W. L. Bragg. The structure 
of diopside CaMg(SiO 3 ) . Z. Krist., 69, 
168-193 (1928). 

Warren, C. H., and C. Palache. The pegmatites 
of the riebeckite-aigerite granite of Quincy, 
Mass., U.S.A.; their structure, minerals, and 
origin. Proc. Am. Acad. Arts Sci., 47, 1 25— 
168 (1911). 

Washington, H. S. Chemical analyses of igneous 
rocks. U. S. Geol. Survey Prof. Paper 99, 
1201 pp. (1917). 

Wasserburg, G. J., F. J. Pettijohn, and J. Lipson. 
A 40 /K*° ages of micas and feldspars from 
the Glenarm series near Baltimore, Mary- 
land. Science, 126, 355-357 (1957). 

Wilkinson, J. A. Liquid hydrogen sulfide as a 
reaction medium. Chem. Revs., 8, 237-250 

Williams, G. H. Contact metamorphism near 
Peekskill. Am. J. Sci., 36, 259 (1888). 

Willis, B. T. M. An optical method of studying 
the diffraction from imperfect crystals, Part 
III. Proc. Roy. Soc, A, 248, 183-198 (1958). 

Wilson, A. J. C. X-Ray Optics. Methuen Mono- 
graphs, 1949. 

Winchell, H. Honolulu series, Oahu, Hawaii. 
Bull. Geol. Soc. Am., 58, 1-48 (1947). 

Wright, W. I. The composition and occurrence 
of garnets. Am. Mineralogist, 23, 436-449 

Wyart, J., and G. Sabatier. Transformations 
mutuelles des feldspaths alcalins. Reproduc- 
tion du microcline et de l'albite. Bull. soc. 
franc, mineral, et crist., 79, 574-581 (1956). 

Yoder, H. S., Jr. High-low quartz inversion up 
to 10,000 bars. Trans. Am. Geophys. Union, 
31, 827-835 (1950a). 

Yoder, H. S., Jr. Stability relations in grossular- 
ite. /. Geol, 58, 221-253 (1950£). 

Yoder, H. S., Jr. The MgO-Al 2 3 -Si0 2 -H 2 
system and the related metamorphic facies. 
Am. f. Sci., Bowen Vol., 569-627 (1952). 

Yoder, H. S., Jr. Role of water in metamor- 
phism. Geol. Soc. Am. Spec. Paper 62, 505- 
524 (1955). 

Yoder, H. S., Jr., and H. P. Eugster. Phlogo- 
pite synthesis and stability range. Geochim. 
et Cosmochim. Acta, 6, 157-185 (1954). 

Yoder, H. S., Jr., and H. P. Eugster. Synthetic 



and natural muscovites. Geochim. et Cos- 
mochim. Acta, 8, 225-280 (1955). 
Zwart, H. J. Regional metamorphism and re- 

lated granitization in the Valle de Aran 
(Central Pyrenees). Geol. en Mijnbouw, 
N. S. 20, 18-30 (1958). 


Scientific Staff 

Director: P. H. Abelson. 

Retired Associates: H. E. Merwin, Penolo- 
gist; E. G. Zies, Chemist. 

Staff Associate: G. J. F. MacDonald. 

Physical Chemists: F. R. Boyd, T. C. Hoer- 
ing, 1 J. F. Schairer, G. R. Tilton. 

Penologists: F. Chayes, J. W. Greig, H. S. 
Yoder, Jr. 

Geochemists: G. L. Davis, H. P. Eugster, 2 
G. Kullerud. 

Geophysicist: S. P. Clark, Jr. 

Physicist: J. L. England. 

Crystallographer: G. Donnay. 

Guest Investigators: R. S. Bradley, Leeds Uni- 
versity, England; Sarah Deutsch, California 
Institute of Technology; H. R. Gault, Le- 
high University; Emilie Jager, Mineralo- 
gisch-petrographisches Institut, Universitat 
Bern, Switzerland; C. R. McKinney, Cali- 
fornia Institute of Technology; C. C. Pat- 
terson, California Institute of Technology; 
C. R. Robbins, National Bureau of Stand- 
ards; L. T. Silver, California Institute of 
Technology; L. L. Thatcher, U. S. Geologi- 
cal Survey; W. Trooskens, Centre d'etude 
de l'energie nucleaire, Brussels, Belgium. 

Fellows: R. G. Arnold, Princeton Univer- 
sity; 3 H. L. Barnes, Columbia University; 
G. A. Chinner, Cambridge University; 4 
L. A. Clark, McGill University; W. G. 
Ernst, The Johns Hopkins University; H. 
Greenwood, Princeton University; 1 N. Mo- 
rimoto, Tokyo University; P. M. Orville, 

Yale University; E. H. Roseboom, Jr., Har- 
vard University; W. F. Schreyer, Univer- 
sity of Munich; A. C. Turnock, The Johns 
Hopkins University; D. R. Wones, Mas- 
sachusetts Institute of Technology; 3 R. A. 
Yund, University of Illinois (National Sci- 
ence Foundation Fellow). 4 

Operating and Maintenance Staff 

Office Manager: A. D. Singer. 

Accountant: E. T. Orozco. 

Editor and Librarian: Miss D. M. Thomas. 

Stenographer: Miss M. E. Imlay. 

Typists: Miss G. C. Andrews, 5 Miss G. Rado- 

sevich. 6 
Stockroom Assistant: M. L. Kirby. 
Mechanic's Helper: C. Baylor. 
Janitor: M. Ferguson. 
Chief Mechanician: F. A. Rowe. 
Instrument Makers: J. F. Kocmaneck, O. R. 

McClunin, G. E. Speicher, W. H. Suerth. 
Mechanic and Carpenter: E. J. Shipley. 
Electrician: E. C. Huffaker. 
Machinist: J. R. Thomas. 
Apprentice-Machinist: W. R. Reed. 7 
Building Engineer: R. L. Butler. 

1 Appointment 

2 Appointment 

3 Appointment 

4 Appointment 

5 Appointment 

6 Appointment 

7 Appointment 

from January 1, 1959. 
terminated September 30, 1958. 
terminated May 31, 1959. 
from September 1, 1958. 
terminated August 1, 1958. 
from July 2 through August 1, 

terminated February 28, 1959. 


Washington, District of Columbia MERLE A. TUVE, Director 



Introduction 225 

Experimental Geophysics 227 

Radio astronomy 227 

Precise position apparatus 227 

Radio hydrogen 231 

Search for new discrete spectrum 

lines 232 

Visitors' programs 232 

Solar studies 232 

Research in cooperation with the 
Jodrell Bank Experimental Station, 
University of Manchester, Eng- 
land 234 

The earth's crust 234 

Seismic studies 234 

Mineral age measurements 237 

Rock magnetism studies in the Spray 
Quadrangle, Oregon 250 

Theoretical and Statistical Geophysics . . . 253 

Equatorial electrojet 253 

Cosmic-ray investigations 254 

Laboratory Physics 256 

Nuclear physics 256 

Polarized ion source development. . 256 

Cyclotron 259 

Biophysics 259 

Introduction 259 

Ribosomes 260 

Amino acid analogs 294 

Cooperative work on the cerebral 

cortex 299 

Cooperative work with other lab- 
oratories 299 

Operations and Staff 300 

Cooperative work of the Department. 300 

Administration and operation 301 

Bibliography 301 

Personnel 303 

Carnegie Institution of Washington Year Book, 58, 1958-1959 

Plate 1 

Department of Terrestrial Magnetism 

The radio astronomy activities of the Department are being greatly augmented by the installa- 
tion during July 1959 at Derwood, Maryland, of a parabolic reflector 60 feet in diameter mounted 
equatorially to follow astronomical objects. This splendid instrument replaces the 26-foot parabola 
in use at Broad Branch Road since 1953 for studies of the hydrogen clouds between the stars of 
our galaxy. 


Physics is today an exceedingly wide- 
ranging field of study and interest. More- 
over, the techniques and analytical pro- 
cedures that are standard tools of the physi- 
cist have in recent years provided fresh 
avenues of approach for the study of 
hitherto inaccessible problems in many 
fields such as geology and biology, where 
active physical interactions and forces are 
usually not the primary focus of attention. 

Along with these new challenges many 
other special areas usually identified with 
physics have become huge industrial are- 
nas, with thousands of workers and ex- 
penditures in the billions of dollars. Radar 
and electronics, solid-state physics, nuclear 
physics, and even upper-atmosphere studies 
are familiar examples of physics converted 
to technology. 

The surprising feature of this modern 
repetition of the old process of converting 
science into technology is that it now takes 
place before the basic scientific studies have 
ripened to reasonable completion. This 
fact is amply clear to the engineers and 
the administrators, with the result that, 
chiefly on government contracts, large 
sums are currently made available for "di- 
rected research," with insistent backing 
from a technological group and urgent 
pressure for desired answers. There are no 
clear boundaries between technology and 
science in such fields as nuclear energy or 
"space science." Nor are there any finan- 
cial or other criteria for the "value" of any 
specific type of data that may be obtained. 
The result all too often is an expenditure 
of millions or tens of millions of dollars 
in quasi-scientific tests that are in actual 
fact only trial runs for some excessively 
confident group in technology, seeking rec- 
ognition by some hoped-for spectacular 

This has changed the approach to many 
inadequately investigated areas of physics 
from rough trails of research adventure to 

broad avenues of industrial competition. 
These spectacular activities in technology, 
publicized as today's crucial "efforts of the 
scientists," are a remarkable indication or 
demonstration of the general aims of our 
society. These costly efforts may in many 
ways be desirable and worth what they 
cost. We hope so. For many of us, how- 
ever, this change of research in physics 
"from a sport to a business" has made it 
seem rather discouraging and more or less 
pointless for one or two quiet research 
men with only modest funds at their dis- 
posal to remain active in certain areas of 
research that have been made fields of pri- 
mary interest to government agencies. 
Radio propagation studies, certain upper 
atmospheric studies, most areas of nuclear 
physics, many areas of solid-state physics, 
and even some aspects of astronomical re- 
search are rather generally in this category. 

The response of the free investigator 
who values the opportunity for quiet study 
and deep personal engagement with his 
own research materials is to select areas for 
his own work which are not likely to be- 
come the special province of some techno- 
logical group for high-pressure "team re- 
search." He in fact doubts the validity of 
such a term; research for him is a highly 
personal commitment, more closely akin 
to portrait painting than to industrial ac- 
tivity. The opportunity (and necessity) to 
guide and control a dozen or thirty work- 
ers in their daily efforts on a problem, 
even a "scientific" problem, does not strike 
him as a research task, but as a manage- 
ment activity, which, indeed, it is. 

As a result of these considerations, and 
many others similarly rooted in our con- 
victions as to the basic nature of research, 
these reports each year cover a wide range 
of subject matter on studies in progress by 
the physicists here. These reports are only 
factual descriptions of a series of human 
adventures, intensely, even passionately, ex- 




perienced. The idea of an amateur as one 
who loves his subject is no stranger to this 

Many of today's experimental probings 
into atoms or living cells or out to the 
distant reaches of space demand expensive 
specialized equipment. Even a modest 
thrust in some new direction often re- 
quires a major apparatus before much of 
a beginning can be made. Our program in 
radio astronomy, which has grown into a 
major effort during the past six years, has 
some of these characteristics. Antennas of 
impressive dimensions are necessary, for 
two reasons: radio wavelengths are long 
(from a centimeter or two to several 
meters), and the intensities of the radio 
emissions received from space are uni- 
formly very low. 

The frontispiece of this report shows the 
new 60-foot parabolic antenna reflector, on 
an equatorial mounting, which is being 
completed at our Derwood (Maryland) 
Field Station as the report year ends. This 
magnificent new instrument, though 
smaller than the 84-foot units installed at 
several points in the United States and 
Europe during the past three years, is an 
instrument of high precision which gives 
a strong base of support for our work. It 
will be used chiefly to extend the long 
series of measurements of the hydrogen 
clouds in our galaxy, made with our 54- 
channel H-line receiver and a 26-foot pa- 
rabola. Some of the work planned for the 
new 60-foot parabola is complementary to 
that with antenna arrays, nearly 3000 feet 
long, constructed with great precision and 
designed to yield precise positions of radio 
objects in the sky. That work has made 
fruitful progress during the year; repeated 
measurements of known objects (in Cyg- 
nus and Taurus) have demonstrated its 
reliability and precision, with errors less 
than a minute of arc. 

The solar program in radio astronomy 
is another facet of our systematic effort to 
evaluate the character and magnitude of 

the contribution to astrophysical knowl- 
edge that can be expected or hoped to re- 
sult from detailed and imaginative studies 
of the radio emissions received from space. 

The Department has continued its long- 
standing activities in the experimental 
study of the crust of the earth, as a guide 
to the early chapters of earth history, by 
isotope measurements of the ages of Pre- 
cambrian minerals and by studies of the 
thickness and other characteristics of the 
crust using waves from explosions. The 
mineral age program was extended in 
scope this year by studies of samples from 
Europe, Arabia, and South America, with 
growing emphasis on the question whether 
there have been a few isolated periods of 
great orogenic activity separated by long 
quiescent or quiet periods in the earth's 
geological history. The evidence to date 
appears to favor this view. Our seismic 
program for studies of the crust began to 
take on a new dimension during the year 
by the installation, near Arequipa, Peru, 
of the first station of a local earthquake 
network planned for Peru, Bolivia, and 
Chile. The very strong attenuation of ex- 
plosion waves under the Andes ranges and 
the altiplano, observed in our reconnais- 
sance during 1957, led to the present effort 
to study the much more energetic waves 
generated by numerous small earthquakes. 
Present studies are directed toward ascer- 
taining the precision with which these local 
shocks can be located in three dimensions 
and in time. 

Statistical and observatory studies this 
year have again been directed to the ex- 
amination of the "electrojet" in the high 
atmosphere at the magnetic equator. An 
intense and narrow band of electrical cur- 
rent flows overhead there, which varies 
sharply with local time and especially dur- 
ing magnetic storms. The five observa- 
tories set up last year in Peru, as a part of 
our IGY effort, are continued during 1959 
to give a full year of comprehensive rec- 
ords for study. 

A succession of disappointments was met 


in our attempts to achieve a polarized pro- 
ton beam for nuclear experiments. The 
joint project with colleagues at Yale Uni- 
versity, designed to use the familiar atomic 
beam apparatus to select protons with one 
direction of spin, turned out to be much 
more elaborate and involved than antici- 
pated, and it has not yet reached the stage 
where its eventual usefulness to us could 
be subjected to test. More decisive has 
been the failure of a device based on the 
selection of metastable hydrogen states, a 
joint effort with friends and colleagues 
at Johns Hopkins. After nearly 18 months 
of effort seeking a polarized proton beam, 
our nuclear physics group is now gradually 
turning to more conventional studies with 
hydrogen and tritium. 

It will occasion no surprise to readers of 
previous reports of the Department to 
note that much of the most striking and 
fruitful activity of the year past was in our 
biophysics group. Even the sheer bulk of 
the biophysics report, below, displays our 
satisfaction in this work, along with the 
astonishing diversity and extensiveness of 
the studies of the synthesis of proteins and 
nucleic acids. Much of the attention of 
our biophysics group has been directed to 

studies of the ribonucleoprotein particles 
(ribosomes) found in bacterial cell homo- 
genates, which appear to provide the sites 
for the synthesis of protein. Quantitative 
studies of these "templates" for proteins 
and enzymes, and of paths from precursors 
to finished vital products, are both fasci- 
nating and rewarding. It should be noted 
that the ribosomes must re-create their 
own structures in addition to soluble pro- 
teins and other components of the cell. 
The relationships between the DNA of 
chromosomes and the RNA of the ribo- 
somes are a current challenge to the group. 
The analytical procedures we use here, in 
addition to the usual separation procedures 
of biochemistry, involve a great many ex- 
periments using the brief "pulse" tech- 
nique, whereby radioactive tracer atoms 
momentarily replace the usual inactive 
elements in some part of the nutrient solu- 
tion, for example the sulfur (as in methi- 
onine). A complementary series of experi- 
ments investigates the frequency and char- 
acter of "mistakes" that enzymes and tem- 
plates can make in protein synthesis by 
studying the incorporation of amino acid 
analogs into bacterial proteins. 



B. F. Bur\e, J. W. Firor, M. A. Tuve, and 
H. W. Wells 


The long-felt need for accurate positions 
of discrete radio sources has been discussed 
in previous annual reports, and the first 
steps toward increasing our knowledge of 
the positions of the brighter sources have 
been outlined. As described, two long V 
reflectors have been constructed for use at 
a frequency of 405 Mc/sec, each 250 wave- 
lengths long and the pair separated by 750 
wavelengths between centers. Either can 
be used as a single antenna with a fan 
beam %° by 20°, or the two can be used 
together as an interferometer with a fringe 
spacing of 4 minutes of arc. The first V 

reflector was put into operation August 
13, 1958; it has served as a valuable test 
instrument, both for checking the sensi- 
tivity of our receivers and for indicating 
the positional accuracy of the system. The 
second array has not yet been used for 

The measurement technique is closely 
related to a well known astronomical in- 
strument, the transit. Since the arrays 
have high resolution in only one dimen- 
sion, only one parameter can be deter- 
mined by a set of observations made with 
the arrays in a given orientation. Initially, 
the arrays are aligned with their largest 
dimension in an east-west direction. As a 
consequence, right ascension of a radio 
source can be measured accurately, but 



declination is only roughly determined. 
By varying the azimuth of the instrument, 
however, a combination of right ascension 
and declination will be measured from 
which the coordinates of the observed 
radio sources can be deduced. Varying 
the azimuth involves remounting the V 
reflectors on a new set of posts with a 
fresh set of transmission lines, an opera- 
tion that will be carried out after a survey 
of the sky has been completed with the 
present east-west orientation. 

The errors in the radio measurements 
are the same as those that must be con- 
sidered with the optical instrument. Azi- 

the results obtained from measurements 
on known radio sources. Allowance must 
be made for detecting atmospheric effects, 
such as dew, and rejecting records that 
might have been affected. A fourth po- 
tentially serious source of error is the con- 
fusion that arises when an apparent source 
is really a blend of two or more sources. 
This error can only be estimated statisti- 
cally, but our studies have shown that such 
errors are not serious if the number of 
sources being measured is less than an 
average of 1 source every 20 beam widths, 
a limit we do not approach as long as we 
are well away from the Milky Way. Last, 


+ 15-- 
+ IO-- 

+ 5 


20 c 

40 c 

-60 c 


Fig. 1. Difference between optical and radio transit time of the four brightest radio sources. 
Individual and average differences are shown with solid curve indicating expected difference for col- 
limation error of % minute of arc. 

muth and level errors are determined by 
accuracy of construction of the instrument. 
The geometry of the arrays allows ac- 
curate control of these errors, which are 
believed to be less than 10 seconds of arc 
for each array and less than 5 seconds of 
arc for the pair as an interferometer. The 
collimation error, being dependent on the 
electrical symmetry of the transmission 
lines feeding the array, is harder to con- 
trol. A phase shift in the lines feeding 
one-half of the array shifts the collimation 
plane by an angle equal to the magnitude 
of the relative phase shift, a quantity diffi- 
cult to measure accurately. We could con- 
trol this error only by making the feeder 
system mechanically symmetrical through- 
out, checking the radiofrequency imped- 
ance at each branch point, and checking 

there is the measurement error itself, aris- 
ing from the accuracy with which an ac- 
tual time measurement can be made. This 
is dependent largely on the signal-to-noise 
ratio, since high accuracy of timekeeping 
can be achieved with modern crystal os- 

Observations of the brightest radio 
sources, which have formed the first phase 
of our program, have yielded a measure- 
ment of the collimation error for the V 
reflector now in operation. The sources 
used were the four brightest discrete 
sources Cygnus A, Cassiopeia A, Taurus 
A, and Virgo A. Figure 1 shows the dif- 
ference in right ascension between our 
observed radio values and the optical posi- 
tions as given by Baade and Minkowski. 
The individual observations are plotted, 


and the average is shown by a vertical bar. 
The curve shown is the error to be ex- 
pected from a collimation error of % min- 
ute of arc, from which it can be seen that 
the difference remaining between the radio 
and optical positions after correcting for 
this error is less than % minute of arc (1 
second of time at the equator) . It appears, 
therefore, that the centers of radio and 
optical emission for the four brightest 
radio sources are closely coincident, in 
right ascension at least. The scatter in 
individual observations is caused partly by 
reading error, though refraction, probably 
in the ionosphere, may contribute to it 

Milky Way at least nine sources have been 
observed. Three of these, Cygnus A, Virgo 
A (M 87), and IAU 16NOA, are well 
known. Of the remainder, all are weak 
and two are definitely of angular extent 
greater than our beam width { X M°). When 
the second V reflector is put into operation 
during the summer of 1959, the stable 
baseline of the interferometer record and 
the increased area of the interferometer 
combination should increase this number 
by a considerable factor. Recent improve- 
ments in receiver technique, notably para- 
metric amplifiers, are being investigated; 
if practicable, they should improve our 
signal-to-noise ratio still further. 

_*j|«_ 1/2 power 
beam width 

I9 n 30 

I8 h 30 m 

Fig. 2. Scan of Milky Way at 8= +10°, showing the rise due to the galactic background with 
three well defined sources superimposed. 

The number of sources that we can 
expect to measure is a question of con- 
siderable interest. Confusion sets an up- 
per limit of approximately four hundred 
sources with 90 per cent probability that 
the apparent position will be shifted less 
than ! /4 minute of arc by blending with 
a weaker source, but it is clear from our 
preliminary survey that we cannot ap- 
proach this limit with present receivers. 
We have used a Dicke-type comparison 
receiver, comparing the antenna tempera- 
ture seen at the terminals of the V reflector 
with the temperature of a small antenna 
looking at the vicinity of the north celes- 
tial pole. The front end of the receiver 
was a conventional grounded-grid pre- 
amplifier with a noise figure of about 4. 
With this equipment, approximately one- 
third of the visible sky has been surveyed, 
and in the sections well removed from the 

In the vicinity of the Milky Way, there 
is a profusion of bright radio sources, so 
that the problem is resolution rather than 
detection. Little improvement can be ex- 
pected from more sensitive receivers, and 
definite conclusions can be drawn from 
the present observations. The principal 
features of the low-latitude observations 
are illustrated by figure 2, which is taken 
from an average of several made at decli- 
nation + 10°. In galactic coordinates, the 
region covered between half-power points 
of the antenna beam stretches from /=0° 
to /=20°. Several discrete sources can be 
seen superimposed upon a continuous 
background. The background results from 
the disk component of the galactic radia- 
tion, a prominent ridge approximately 3° 
wide, roughly symmetrical about the galac- 
tic plane, reaching its greatest intensity in 
the vicinity of the galactic center and 


fading out toward the anticenter. Three 
sources are indicated on the record, two 
of which are clearly wider than the an- 
tenna beam. The corresponding right as- 
censions have been plotted in figure 3, 
superimposed upon a map prepared in 
Leiden from Westerhout's survey of the 
galaxy at 1390 Mc/sec, with a pencil beam 
which measured 45 minutes of arc to half- 
power points. Despite their having greater 
resolution in declination and less resolu- 




Fig. 3. Section of Westerhout's 21 -cm con- 
tinuum survey, with right ascension of three 
prominent sources at 405 Mc/s indicated by 
vertical lines. The 21-cm brightness is shown in 
the form of equally spaced contours plotted on 
celestial coordinates. 

tion in right ascension, comparison of the 
results at the two frequencies yields some 
information about the nature of the more 
prominent sources. The sources marked 
18E and 19C are clearly extended, with in- 
tegrated intensities that are slightly lower 
at our frequency of 405 Mc/sec. 

The implication is that these sources are 
ionized hydrogen regions, whose spectral 
characteristics are known to exhibit such 
behavior. On the other hand, source 19B 
shows no evidence of extension, and is 

slightly more intense at the lower fre- 
quency, the ratio being about the same as 
for the well known Crab Nebula, which 
is a nonthermal emitter of radio noise. 
Other features can be seen in figure 2 
which are undoubtedly correlated with 
features in the 1390 Mc/sec map, and are 
probably associated with thermal emission 
from ionized regions, but we have concen- 
trated attention upon the most clearly de- 
fined sources. Unfortunately, heavy ob- 
scuration by interstellar dust clouds ham- 
pers optical identification in this region, 
but it is hoped that our interferometer 
observations during the coming year (to 
establish order of magnitude of angular 
size and hence brightness temperature), 
combined with radio data at other fre- 
quencies and optical correlations when 
possible, will shed some light upon the 
types of radio emitters existing in our 
galaxy. At the time of writing, about 40 
per cent of the galactic circle had been 
surveyed and a list of seventeen bright fea- 
tures compiled. Since the observed region 
contains many of the brightest sources, it 
is estimated that perhaps thirty bright fea- 
tures close to the Milky Way will be 
observable, from which the space density 
of galactic sources in the vicinity of the 
sun can be estimated. 

A galactic source of particular interest 
at present is the celebrated Sagittarius 
source, Sgr A, often identified with the 
center of our galaxy since it lies in that 
direction. Early in the spring of 1959, 3-cm 
observations at the National Radio As- 
tronomy Observatory indicated that the 
source was complex, and within weeks we 
were able to complement these observa- 
tions with our own at 405 Mc/sec. A pre- 
liminary curve is shown in figure 4, and 
compared with that obtained by Drake at 
9000 Mc/sec. When allowance is made 
for the difference in beam width, it can 
be seen that the two curves are remark- 
ably similar. There is an uncertainty in 
the zero level of the 405 Mc/sec curve, 
because the underlying galactic back- 


ground had to be removed, but the rela- 
tive intensities at the two frequencies sug- 
gest a nonthermal spectrum. A unique 
separation of the principal complex can- 
not be made, but a possible interpretation 
is a broad source, with a source of smaller 
angular diameter superimposed off-center. 
Either one or both of the sources may still 
be associated with the galactic center, but 
the relationship is not yet clear. Obscura- 
tion by dust clouds hampers optical obser- 
vations. (B. F. B., J. W. F.) 

are installing a 60-foot parabola on an 
equatorial mounting at the Derwood Field 
Station of our Department, 16 miles north 
of Washington. This magnificent new 
equipment is a direct example, detailed for 
fabrication by Blaw-Knox, of the design 
worked out here by the late Dr. Howard 
Tatel. It incorporates primary structural 
elements based on tetrahedrons and the 
use of large-diameter gears to give ac- 
curacy in setting angular positions and to 
minimize gear stresses under windstorm 

1/2 -power beom widths 

405 Mc/s 
8000 Mc/s 

Fig. 4. Comparison of scan through the "galactic center" at 405 Mc/s with Drake's 3-cm results 
for the same region. 


The 54-channel hydrogen-line receiver 
was used with the 8-meter Wiirzburg 
parabola again during the entire report 
year to obtain measurements of atomic 
hydrogen density (apparent temperature) 
and Doppler shift for 470 additional points 
in the sky near the galactic plane. These 
observations continue and extend the sur- 
vey discussed in our report last year, as 
described to the Paris Conference in Au- 
gust 1958. The diffuse values obtained 
with this small parabola, which averages 
over an area of sky roughly 2° by 3°, 
serve to outline the interesting regions 
selected for more detailed study during 
the coming year with our new 60-foot 
parabola (see Frontispiece). 

During the summer months of 1959 the 
Blaw-Knox Company and the Radio Con- 
struction Company, both of Pittsburgh, 

conditions. Control of construction ac- 
curacy and acceptance testing for the De- 
partment are under the supervision of 
Mr. Everett Ecklund, of our staff. The 
first use of the equipment on radio objects 
in the sky is expected in the early autumn 
of 1959. 

The 54-channel hydrogen receiver has 
been modified twice during this year. Both 
times extensive changes were made in 
the local oscillator system at 1393 and 
1391 Mc/sec. A set of six 21-Mc/sec crys- 
tals with doublers and triplers to give 
crystal-controlled heterodyne frequencies 
for measuring Doppler velocities of the 
hydrogen did not prove satisfactory. Small 
changes of tuning condensers or in envi- 
ronmental conditions would shift the 1393 
Mc/sec frequency by as much as 10 or 
even 20 kc/sec, thus shifting the apparent 
hydrogen velocity by 2 to 4 km/sec. As 



detailed monitoring of these drifts proved 
troublesome, a new basic oscillator was 
designed. In this latest version the output 
of a calibrated variable signal generator 
near 2.6 Mc/sec is mixed with a 22.7 
Mc/sec frequency from a temperature- 
controlled crystal to give 25.3 Mc/sec. This 
is again mixed with the quadrupled crys- 
tal frequency at 90.8 Mc/sec to give a basic 
oscillator frequency at 116.1 Mc which is 
under the control of the variable signal 
generator near 2.6 Mc/sec. This arrange- 
ment has the special advantage that an 
error of 100 cycles/sec at 2.6 Mc/sec re- 
mains an error of 100 cycles/sec at 116 
Mc/sec instead of being multiplied 40 
times. The 116 Mc/sec output is tripled 
and quadrupled to give the required 1393 
Mc/sec which is mixed with the 1420.4 
Mc/sec hydrogen signal to give the re- 
quired 27 Mc/sec intermediate frequency 
used in the set. This new local oscillator 
system has proved highly satisfactory in 
two months of service. Shifting the posi- 
tion of zero-Doppler hydrogen (at rest) 
to the upper or lower ranks of the 54 chan- 
nels is readily done, since the variable 
frequency signal generator is located at 
the recording position. No spurious ef- 
fects attributable to pickup or "cross talk" 
have been observed. (M. A. J '., B. F. B., 
E. T. E.) 


In a report at the IAU Moscow meeting 
it was suggested that transitions in hydro- 
gen between levels of very high quantum 
number might be observable. The n — 
167-166 transition falls at a frequency of 
1424.632 Mc/sec, and it was an easy matter 
to retune the local oscillators of the 21-cm 
receiver to look for this line. The best 
chances for observing it would be expected 
to be found in an H II region, where 
sufficient recombinations are occurring to 
populate the high-« levels. Observations 
of the Orion region, Cygnus X, the galac- 
tic center, and Spica yielded completely 
negative results. The line intensity aver- 

aged over a 20 kc/sec bandwidth was less 
than 1° K in all cases. (B. F. B., J. W. F., 
M. A. T.) 


H. Lawrence Heifer and Malcolm P. 
Savedoflf, of the University of Rochester, 
visited the Department on several occa- 
sions during the year for the purpose of 
making hydrogen-line observations with 
the 8-meter Wiirzburg and multichannel 
receiver. A model of the velocity field of 
the near-by hydrogen is being constructed 
by Heifer, who has shown that the local 
motions of the H I in the solar neighbor- 
hood show symmetry about a pole that is 
surprisingly far removed from the galactic 
pole. The position of the "hydrogen pole" 
coincides most closely with the pole de- 
rived by Shajn from interstellar polariza- 
tion measurements. Since the local galactic 
magnetic field determines the direction of 
interstellar polarization, it would appear 
that the gas motions are also affected. This 
would seem to be a manifestation of the 
influence of the local galactic magnetic 
field on the gas motions, since the inter- 
stellar polarization probably arises from 
alignment of interstellar dust grains in 
the magnetic field within the spiral arms 
of our galaxy. 

In a separate project, it is hoped that 
the total amount of hydrogen in the di- 
rection of the galactic poles can be esti- 
mated. The expected extinction in the 
vacuum ultraviolet beyond the Lyman 
limit depends critically on this quantity, 
knowledge of which will affect the types 
of experiments being planned for astro- 
nomical satellite experiments. 

W. C. Erickson, of the Convair Scien- 
tific Research Laboratory, completed a 
series of observations to supplement the 
high-latitude survey now being prepared 
for publication. 


The long antenna array with which the 
sun is scanned at a frequency of 340 
Mc/sec (wavelength of 88 cm) has been 


in operation for about two years. The 
several hundred scans obtained permit 
some features of the radio sun to be 

The "quiet" sun continues to be an in- 
teresting problem even though the last two 
years have been a period of great activity 
on the sun. It is well known from the 
Australian work that at short wavelengths, 
around 20 cm, the quiet-sun radiation can 
be separated from the temporary bright 
spots even during active periods by super- 
imposing the scans for several weeks and 
drawing the lower envelope to the group 
of curves. At meter wavelengths, how- 
ever, for example at 176 cm, such a pro- 
cedure does not give reproducible results; 
the scans, even though apparently quiet 
as judged from lack of bursts and large 
hot spots, change their form from day to 
day and from month to month. It is like 
trying to gain a picture of the surface of 
a planet by superimposing many photo- 
graphs of the planet taken at different 
times. The 20-cm results could be com- 
pared with the results on the planet Mars, 
where the clouds do sometimes clear and 
parts or all of the surface are visible. The 
accumulation of photographs would there- 
fore finally lead to a full description of the 
surface. The long-wavelength solar re- 
sults, on the other hand, suggest photo- 
graphs of Jupiter — no matter how many 
are added together, all that can be seen is 

Like the 20-cm scans, the 88-cm scans 
can be superimposed to give a fairly repro- 
ducible lower envelope if they are taken 
over a sufficiently long period, say three 
months or more. This lower envelope is 
consistent with that expected at 88-cm 
wavelength from a corona having the 
temperature and density as derived from 
optical eclipse observations. One feature 
suggestive of the meter-wavelength obser- 
vations should be noted. Upon occasion, 
the daily scans will be quiet for a week 
or so, but not identical with the lower 
envelope derived for a long period. Dur- 

ing these quiet periods the day-to-day 
changes are small, and frequently at least 
half of the disk will be completely un- 
changed for several days. That is, it is 
sometimes possible to derive a satisfactory 
lower-envelope scan for a short period, 
and such a scan differs from a scan de- 
rived for a long period. As far as the 
corona is concerned, this observation indi- 
cates that the part of the corona respon- 
sible for the 88-cm radiation can vary in 
its characteristics, not only in localized hot 
spots but also in extensive regions the size 
of the disk itself. 

The study of the localized bright spots 
comparing the work of several labora- 
tories at different wavelengths, mentioned 
in the previous report, has been com- 
pleted, mainly through the efforts of D. O. 
Mathewson, of the Sydney group. It shows 
that a model of a slowly varying solar 
active region can be constructed from the 
radio measurements alone, if they are 
available at a variety of wavelengths and 
if the period is such that the longer 
wavelengths are not too perturbed by non- 
thermal radiation. The nonthermal radi- 
ation itself was not included in Mathew- 
son's study. 

An attempt was made at the Depart- 
ment to study all the nonthermal or "noise 
storm" events that have occurred in the 
last two years and that were observed both 
at the Department at 88-cm wavelength 
and in France at 176-cm wavelength. It 
was concluded that most noise storms oc- 
cur at both wavelengths with a ratio of 
intensity, at peak intensity, of less than 
10, and with the longer wavelength giving 
the larger flux. If the noise storm begins 
near the center of the solar disk, the 
shorter-wavelength storm develops about 
a day ahead of the longer-wavelength one. 
The positions of the storm source at the 
two wavelengths in general agree, but 
with the longer-wavelength storm always 
a bit farther from the center of the disk 
than the 88-cm position. 

Taken together, these results imply that 



the condition resulting in a noise storm 
proceeds upward through the solar atmos- 
phere and causes first the 88-cm storm and 
then, later and higher, the 176-cm storm. 
For many days when a noise storm was 
observed above one limb of the sun, the 
prominence photographs of the sun taken 
in the light of the H a at the High Altitude 
Observatory of the University of Colorado 
were examined. No particular correlation 
of noise storms and optical features was 
noted. It may well be that the change in 
the corona that results in a noise storm is 
not sufficiently drastic to produce optically 
observable phenomena except under un- 
usual circumstances or with new observing 
techniques. (/. W. F., B. F. B.) 





The research activities initiated in Sep- 
tember 1958 at Jodrell Bank after discus- 
sions with Professor A. C. B. Lovell and 
his staff have continued in two produc- 
tive areas : radio astronomy at frequencies 
below 30 Mc/sec, and unusual radiofre- 
quency emission and absorption at 80 

One objective in the low-frequency work 
has been to establish the relative intensities 
of several principal radio sources and to 
resolve the large discrepancies in the re- 
sults of the very few investigators who 
have attacked this problem. Knowledge 
of characteristics of radio stars at these 
frequencies is a prerequisite to the under- 
standing of generation processes and ef- 
fects of galactic or extragalactic absorption. 
In this work the world's largest steerable 
radio telescope has been used extensively, 
either alone for total power measurements 
or in conjunction with other aerials to 
form an interferometer. The preliminary 
results are extremely encouraging. Sev- 
eral successful series of measurements have 
been made to frequencies as low as 16.8 
Mc/sec. The ratio of Cygnus to Cassiopeia 
appears to be relatively unchanged at 

about 0.5 over the entire frequency span. 
Sporadic emission and absorption at 80 
Mc/sec is a rather rare effect and is quite 
localized in the outer atmosphere. Obser- 
vations have established that noise-type 
emission in one sector of the sky is often 
associated with absorption of the galactic 
background in an adjacent sector. The 
physical processes by which radiofrequency 
noise can be generated in our atmosphere 
are of special interest and significance in 
view of the rapidly increasing knowledge 
of radiation belts and "trapped" high- 
velocity particles in the earth's exosphere. 
It is believed that the observed phenomena 
may be the result of impact of high- 
velocity particles (soft cosmic rays?) gen- 
erating "noise" by a synchrotron process 
in the outer atmosphere and producing 
absorption as the result of ionization when 
the confined rays or streams hit the lower 
ionosphere. One technical report on this 
subject has been published, and another 
is in press. (H. W. W.) 


L. T. AUrich, M. N. Bass, 1 M. A. Tuve, and 
G. W. Wetherill 


Our 1957 reconnaissance of the high 
Andes in Peru, Bolivia, and Chile, dis- 
cussed in last year's report, showed that 
explosion waves could be recorded up and 
down the west flank of the Andes but 
that extreme attenuation, without prece- 
dent elsewhere, prevented observation of 
waves from rather large open-pit mine 
shots (60 tons of explosive) beyond 230 
kilometers across the mountain ranges and 
the altiplano. Observations were made at 
sites so quiet that undulations from the 
shot as slight as 1/10,000 of a wavelength 
of light (or one-half an angstrom unit) 
would have been reliably detected. 

The encouraging records on the western 
flank, which gave indications of irregular 
but definitely measurable depths to the 
M discontinuity which is presumed to 

1 Carnegie Institution Fellow; from North- 
western University. 


mark the bottom of the earth's crust (top 
of the earth's mantle), together with mys- 
terious attenuation across the mountains, 
led to discussions and plans for future 
observations of local earthquakes in these 
regions while our IGY expedition was still 
in Chile. Local earthquakes in the Andes 
are often of rather impressive magnitude, 
and are unequivocally observed across the 
mountains as well as along the flanks. The 
practical problems to be met in gathering 
the required earthquake records are those 
of manpower and logistics. 

The many breakdowns of our trucks on 
the Andes roads and our troubles on the 
back trails where quiet sites are found did 
not encourage the idea of installing stand- 
ard earthquake instruments in light-tight 
shelters or vaults, with photographic rec- 
ords to be changed daily. The network of 
12 to 16 stations required for such proce- 
dures would tax the resources of a large 
establishment, whereas our work is always 
planned on a highly personal basis, to be 
carried out by a few individuals deeply 
interested in the questions under study. 
While still in the Andes we therefore 
planned a set of seismic recorders with 
electronic instrumentation which would 
require attention only once a week. Dur- 
ing this year the prototype sets of these 
quake recorders have been constructed 
here and the first sets have been installed 
by Dr. A. Rodriguez B. in the Andes near 
Arequipa, Peru. Dr. Rodriguez, Professor 
of Geophysics at the San Agustin Uni- 
versity in Arequipa, spent two months in 
Washington helping us test out the in- 
strument design, and he has arranged to 
operate a network of five stations near 
Arequipa as a major activity of the Insti- 
tute of Geophysics of his University. Pen 
and ink records are made on a drum 45 
inches in diameter, the trace spiraling 12 
turns per inch on paper 36 inches wide. 
This drum accepts nearly 8 days of re- 
cording at 2 mm per second; chronometer 
marks are made at 1-minute intervals, and 
the equipment requires only one fresh 

battery at weekly or bi-weekly intervals. 
Pen friction makes the record discriminate 
somewhat against small wave amplitudes, 
but for the study of local quakes out to 
500 or 800 kilometers this fault can be 
tolerated. Preliminary arrangements have 
been made with University groups in 
Bolivia and Chile to extend the network 
out to roughly 500 kilometers after a few 
months of test operation in Peru. A few 
weeks of records already indicate that local 
earthquakes of suitable intensity in the 
vicinity of Arequipa are rather frequent; 
the technique for pinpointing their loca- 
tions and depths, however, may require 
some relocation of the instruments from 
the first sites selected. 

In addition to the earthquake observa- 
tions it is clear from our 1957 observations 
on the large shots in the new open-pit 
copper mine at Toquepala, Peru, that a 
systematic and rather detailed set of obser- 
vations of these explosion waves on the 
west flank of the Andes will give clear 
evidence about the wave velocities and 
thicknesses of the subsurface structures 
down to depths of 40 or 60 kilometers. 
A similar detailed and therefore lengthy 
program of field observations on open-pit 
mine explosions in New Mexico, Arizona, 
Utah, and Wyoming would be expected 
also to yield definitive information on the 
structures beneath our own great western 
plateaus, on the basis of our reconnaissance 
there in 1954. Along with the unattended 
earthquake recorders for the Andes we 
have therefore attempted to work out a 
relatively portable equipment for briefer 
periods of attended recording for observ- 
ing explosion waves. This equipment is 
undergoing field tests in Montana and 
Wyoming during the summer of 1959. A 
difficulty that reduces the amount of usable 
data from explosions is the necessity for 
accepting sites which are relatively noisy, 
owing to local activity. Observations are 
needed every few kilometers, on various 
radiants from the quarry or open-pit mine, 
and local noise of large amplitude on a 



spiral record spoils several traces on each 
side of the noisy one, but the recorder 
must operate continuously if the shot times 
are not controlled. Noise precisely at the 
instant of arrival of the shot waves is 
relatively infrequent, in our experience. 

This technique of using many unat- 
tended recorders for explosion-wave stud- 
ies may require major changes, but if it 
can be developed to a satisfactory stage 
it should enable us to obtain detailed 
coverage of interesting regions that would 
be prohibitively expensive on our old pat- 
tern of time-scheduled shots, using a man 
at each recording site alerted by radio to 
start the recorder 2 minutes before the 

During this year rather vigorous atten- 
tion has been given by other investiga- 
tors, especially those concerned with phase 
velocity measurements on surface waves 
from earthquakes, to our unexpected find- 
ings in 1954 on the Colorado Plateau and 
in 1955 in Alaska and the Yukon, where 
rather deep "roots" of the lighter crustal 
rocks under the high mountain plateaus 
might be expected to extend downward 
into the heavier rocks of the mantle. The 
wave patterns we observed in the moun- 
tains were not markedly different from 
those along the Atlantic seaboard and 
elsewhere at low elevations, with indicated 
depths to the M discontinuity in the range 
29 to 36 kilometers in different localities. 
These figures become actual depths of 34 
to perhaps 43 kilometers when certain 
estimated corrections are made (for slow 
surface rocks and for a possible increase 
of velocity with depth). 

Much of the point of these discussions 
has evaporated, however, by reason of a 
quiet abandonment by others of the old 
view (current until very recently, and 
still held by many workers) that there 
exists only a small or modest contrast in 
density between the lower crustal rocks 
and the upper mantle rocks. The new 
position taken by several of these investi- 

gators visualizes a large density contrast, 
changing from 2.5 or 2.6 for the crust to 
3.27 for the upper mantle. Even for a 
mountain region like the Colorado Pla- 
teau, averaging 6600 feet in elevation, this 
large contrast in densities would predict 
a root of only 7 or 8 kilometers, which is 
not impressively greater than the errors in 
measuring depth to the M discontinuity 
by the various methods. The old estimates 
of rock density, contrasting 2.85 or even 
2.93 in the crust with 3.10 to 3.15 in the 
mantle, would predict roots of 18 to 34 
kilometers, which should show up con- 
spicuously in the field observations. Grav- 
ity surveys and calculations are the basis 
for density estimates, but the gravity cal- 
culations indicating the large density con- 
trast are chiefly derived from oceanic re- 
gions distant from and unrelated to the 
mountain structures under study. We be- 
lieve that discussion of these contrasting 
views is highly desirable, as it exhibits 
many of the unknowns and guesses pres- 
ent in much of our thinking about earth 
structures which have not been examined 
by well bores, and the interest aroused by 
the discussions may lead to intensive stud- 
ies of a few special and limited regions by 
more than one technique. 

As the report year closed, preparations 
were largely complete for sending an expe- 
dition with three trucks and six complete 
seismic recordings to join a geophysics 
group under Dr. Robert Meyer, of the 
University of Wisconsin, for shot measure- 
ments in Montana and Wyoming. Our 
observers will test out several procedures 
for simplified observing, and we welcome 
the opportunity to learn the Wisconsin 
procedures for loading and exploding their 
own shots. Our practice has been to call 
on the Navy or the Coast Guard for the 
shooting operations, which greatly restricts 
the possible sites, or to utilize the regu- 
lar operating explosives of large open-pit 
mines or quarries. 



L. T. Aldrich, G. W. Wetherill, M. N. Bass, 1 

W . Compston? G. L. Davis, 3 and 

G. R. Tilton 3 


During the past few years, much of the 
effort of the various laboratories engaged 
in geochronology has been directed toward 
finding the location in time and space of 
the great Precambrian mountain chains, 
or orogenic belts. As described in previous 
annual reports, the most extensive work 
of this kind has been done on the North 
American continent. It has been found 
that in the eastern United States the Paleo- 
zoic Appalachian orogenic belt is approxi- 
mately paralleled by an older system about 
1000 million years of age. In the western 
United States, data have been found indi- 
cating another zone of orogeny which 
took place 1350 million years ago, and 
the existence of 2600- and 1750-million- 
year orogenic events has also been dem- 
onstrated. Similar, but less extensive, data 
now exist for all the continents. These 
results suggest that there have been cer- 
tain periods of time during earth history 
when, on a world-wide scale, mountains 
were formed, great thicknesses of sedi- 
ments were deformed and reconstituted 
by metamorphism, and large-scale igneous 
intrusions took place. Although it may 
be tempting to draw such a far-reaching 
conclusion from the data, it is probably 
not warranted at the present time. The 
essential difficulty is that, even though an 
approximately correct result is certainly 
obtained for the age of the Precambrian 
orogenies, the inherent resolution of the 
dating methods is not known. When ages 
of, for example, 1750 and 1600 million 
years are obtained on neighboring rocks 
it cannot be said whether they represent 
two distinct orogenic events or two dis- 
tinct phases of the same orogeny if they 
are simply two dates within a long and 

2 Carnegie Institution Fellow; from University 
of Western Australia. 

3 Geophysical Laboratory, Carnegie Institution 
of Washington. 

complex orogenic period, or whether they 
merely represent the failure of one or both 
samples to form a chemically closed sys- 
tem. Progress in relating the results of 
age determinations to the major problems 
of earth history now depends on a deeper 
understanding of the nature and duration 
of an orogenic process and on a more 
detailed understanding of the way in 
which the various aspects of an orogeny 
are recorded in the age of minerals. 

Because of these considerations the work 
done here during the past year has pro- 
ceeded in two directions. First, the gen- 
eral, world-wide study of the approximate 
age of Precambrian igneous and meta- 
morphic rocks has been extended. The 
previous work of this kind in the United 
States and Canada has been continued, and 
new measurements have been made on 
samples from South America, Arabia, and 
northern Europe. These measurements 
contribute to the filling-in of the world- 
wide picture of the history of orogeny 
during geologic time that is taking form 
as a result of the work of a number of 
institutions. At the same time, the some- 
times ambiguous results of these meas- 
urements have underscored the difficulty 
of resolving these various Precambrian 
events. The other direction of our work 
has led to a more detailed study of a 
single orogeny, the Appalachian. The po- 
sition of this orogeny in the geologic time 
scale is very favorable for such a study. 
Events connected with this orogeny took 
place over an interval of time from about 
450 to about 250 million years ago. These 
ages are sufficiently old to permit signifi- 
cant measurements to be made by all the 
major dating methods, but are young 
enough so that the total time span is very 
far outside the uncertainty caused by the 
experimental error in the chemical anal- 
yses. This more detailed study has shown 
that the spread in ages found on geo- 
logically related rocks in the Precambrian 
is apparently a real feature of the orogenic 
process and is more dramatically revealed 



in these younger rocks where the spread 
represents a much greater fraction of the 
total age of the rocks. Because the results 
of the work in the Appalachian region 
shed light on the results of the Precam- 
brian studies, these measurements will be 
discussed first. 

Mineral Ages in the Appalachian 
Oro genie Belt 

As generally pictured, an orogeny usu- 
ally begins with a great thickness of sedi- 
mentary rocks commonly deposited upon 
a basement of older metamorphic and 
igneous rocks. As a consequence of poorly 
understood processes within the earth's 
crust these deeply buried sedimentary 
rocks are deformed by folding and shear- 
ing and are raised to high temperatures. 
The combination of high pressure and 
temperature, probably assisted and cata- 
lyzed by aqueous solutions, causes the 
sedimentary rocks to be recrystallized and 
become metamorphic rocks. In this process 
the older basement may or may not be de- 
stroyed and reconstituted as a younger 
rock. At about the same time magmas, 
sufficiently liquid to be mobile, intrude all 
these rocks, forming granites and other 
types of igneous rocks. The magmas may 
arise from fusion of the sedimentary and 
basement rocks or may represent fresh 
material derived from deeper regions in 
the earth. After a time this active process 
ceases; the material cools and becomes 
solid igneous rock. Following the orogeny, 
probably as a consequence of isostatic ad- 
justment within the earth's crust, this com- 
plex of igneous and metamorphic rocks is 
raised above sea level and exposed by 
the forces of erosion. The time scale of 
these events is almost entirely unknown, 
whether a few hundred thousand or many 
millions of years are required for the 
material to cool to ordinary temperatures, 
whether the processes described above op- 
erate continuously over hundreds of mil- 
lions of years within one orogenic period 
or only as discrete events at one or more 

times. The measurements reported here 
represent an effort to learn something 
about such questions. Because of the mag- 
nitude and complexity of the problem, 
progress will necessarily be slow. 

One early result of this study was the 
discovery that it is still possible to identify 
the older basement gneisses along a large 
part of the Appalachian belt. In agreement 
with similar measurements made at Co- 
lumbia University these older rocks seem 
to have an age of about 1000 million years 
and probably represent an extension of the 
Grenville orogeny in the Canadian Shield. 
The record of their exact age, however, 
has been somewhat obscured by the effects 
of the later orogeny. These rocks tend to 
occur on the western margins of the zone 
of most intense metamorphism, the Pied- 
mont. Perhaps their equivalents farther 
east were destroyed by the intense meta- 
morphism and igneous activity. The re- 
sults of the age measurements on these 
basement gneisses are shown in table 1. 
The K-A ages on the biotites are usually 
younger than the Rb-Sr ages, probably 
owing to thermal loss of argon. The zir- 
con ages, where concordant, indicate a 
similar age, about 1100 million years. Even 
when the zircon ages are not concordant, 
the high Pb 207 /Pb 206 values suggest a Pre- 
cambrian age. The biotite ages on the 
samples of Baltimore gneiss from Mary- 
land and Pennsylvania indicate that these 
micas have been severely altered, or even 
re-formed, by the Paleozoic Appalachian 
events. The biotites from the Maryland 
samples give the lower age, whereas the 
Pennsylvania samples show a very com- 
plex spread, the discordance of which 
probably means that none of their biotite 
ages can be simply interpreted as the time 
of a geological event. The high K-A ages 
relative to the Rb-Sr ages are a great 
puzzle; they may represent absorption of 
radiogenic argon released from other min- 
erals during the heating accompanying the 
Appalachian orogeny. 

The most extensive work on the younger 

TABLE 1. Mineral Ages of Appalachian Precambrian Gneisses 



Bear Mountain, N. Y. 
Canada Hill gneiss 

Storm King granite 

Hibernia, N. J. 
gneiss (dark) 
gneiss (light) 

Devault, Pa. 
Baltimore gneiss 
feldspar altered 
feldspar fresh 

Conshohocken, Pa. 
Baltimore gneiss 

Phoenix, Md. 
Baltimore gneiss 

Towson, Md. 
Baltimore gneiss 

Woodstock, Md. 
Baltimore gneiss 

Loch Raven, Md. 
Hartley augen-gneiss 

Shenandoah National Park, Va. 
Mary's Rock Tunnel gneiss 

Deyton Bend, N. C. 
cranberry gneiss 

Pardee Point, Tenn. 
cranberry gneiss 

Nowhere Ridge, Tenn. 
cranberry gneiss 

Mortimer, N. C, gneiss 
Crossnore, N. C, granite gneiss 
Parkersburg, W. Va. 

Age, million years 


K U 238 U 235 
A Pb 206 Pb 207 

Pb 20T 
Pb 206 

Th 232 
Pb 208 
































































1140 1150 1170 

960 990 1060 


1010 1045 1120 

960 1020 1120 

1040 1070 1120 




1070 1100 1150 

1080 1140 1270 

670 735 940 

800 860 1020 
690 720 800 





All ages used in this report have been calculated with the following decay constants: 

Rb 87 : l^XlO-^yr- 1 

K 40 : X^ = 5.85X10- 11 yr- 1 

A[3 = 4.72X 10- 10 yr- x 

U 238 : 1.54 X lO-^yr" 1 

U 235 : 9.71 X lO" 10 yr- 1 

Th 232 : 4.99X10- 11 yr- 1 




rocks has been done in the vicinity of Bal- 
timore. Here, as elsewhere in the Appala- 
chian Piedmont, younger granites and peg- 
matites of Appalachian age intrude the 
older gneisses represented by the Balti- 
more gneiss, as well as the great sequence 
of metasedimentary rocks known through- 
out the central Appalachians as the Glen- 
arm series. The results of these measure- 
ments are shown in table 2. The K-A ages 

stated is that the effects of the orogeny on 
these rocks appear to have extended over 
a period of at least 150 million years. This 
suggests that similar effects observed in 
Precambrian rocks are probably real, even 
though the time interval involved is 
smaller relative to the analytical uncer- 
tainty. Any further statements are highly 
speculative, but it may be that the 450- 
million-year figure represents the actual 

TABLE 2. Mineral Ages of Younger Rocks from the Washington-Baltimore Area 


Age, million years 






Pb 207 

'J , J 1 232 



Pb 206 

Pb 207 


Pb 208 

Woodstock granite 









Ellicott City granite 









Guilford granite 



Kensington granite gneiss 









Daniels pegmatite 




Loch Raven pegmatite 




Tovvson pegmatite 




Henryton pegmatite 




Baltimore pegmatite 




Guilford pegmatite 1 



Guilford pegmatite 2 



Ellicott City pegmatite 




* K-A ages from the paper of Wasserburg, Pettijohn, and Lipson, Science, 126, 355 (1957). Splits 
of the samples used by these authors were used for the Rb-Sr determinations. 

on some of the samples are the work of 
Wasserburg, Lipson, and Pettijohn. In 
these cases the Rb-Sr measurements have 
been made on splits of the same samples 
used by these authors. All the Rb-Sr ages 
on the samples of biotite as well as most of 
the K-A ages on the biotites and mus- 
covites indicate an age of about 300 million 

It would be tempting to interpret this 
as the time of principal intrusion of these 
rocks were it not for the consistently high 
Rb-Sr ages obtained on the coarse-grained 
feldspars and micas from the pegmatites. 
These ages rule out any such simple in- 
terpretation, and all that can be definitely 

time of the pegmatite and granite crys- 
tallization and that the feldspars and mus- 
covites indicating this age formed a closed 
system with respect to rubidium and stron- 
tium since that time. The position of these 
rocks in the geothermal gradient may have 
been such that for an interval of 150 mil- 
lion years argon and strontium diffused 
from the biotites while only argon diffused 
from the feldspars and muscovite. On the 
other hand, the 300-million-year ages may 
result from a subsequent minor event that 
did not cause the muscovites and feldspars 
to lose strontium. From quantitative con- 
sideration of the variable Rb/Sr ratios in 
the various minerals it does not seem likely 



Very Old Rocks of the Southern Canadian Shield 


Age, million years 

Rb-Sr K-A 

1. Pegmatite, Silverleaf Mine, Manitoba Lepidolite 

2. Granite, east of Kenora, Ontario Biotite 

3. Inclusion in Vermilion granite Biotite 

4. Granite, gneiss, Sioux Lookout, Ontario Biotite 

5. Pegmatite, Hearst, Ontario Lepidolite 

6. Granite, Keefer Township, Ontario Biotite 

7. Inclusion, Round Lake batholith, 

Kirkland Lake, Ontario Biotite 

8. Granite, Round Lake batholith, 

Kirkland Lake, Ontario Biotite 

9. Biotite-rich rock, Cobalt, Ontario Biotite 







that the pegmatite ages are erroneously 
high because of the presence of extra radio- 
genic strontium at the time of their min- 
eralization. This possibility can be checked 
experimentally by measuring the isotopic 
composition of strontium in calcium min- 
erals such as apatite which may be 
formed in a suitable pegmatite. It should 
also be possible to test these other hy- 
potheses by suitable measurements in care- 
fully selected areas. The search for the 
proper samples will be difficult, but the 
value of the information potentially ob- 
tainable makes the necessary effort seem 
worth while. 

Southern Canadian Shield 

Ontario. A compilation of data ob- 
tained here on the great 2600-million-year- 
old region is given in table 3. New meas- 
urements made this year confirm the pre- 
vious discussion of this striking feature of 
North American geochronology. 

These granites and pegmatites are prob- 
ably not the oldest rocks in this region. 
Where the field relationships can be ob- 
served, these rocks are always younger 
than metasedimentary and volcanic rocks. 
Some of these older sedimentary rocks con- 
tain boulders of still older granites. It has 
frequently been suggested that the dating 
of these boulders might result in finding 
the oldest rocks known. In collaboration 
with G. J. Wasserburg, rubidium and 

strontium measurements were made on 
three such boulders from the vicinity of 
Sioux Lookout. These samples were taken 
from the Abram Series conglomerate at 
the north corner of Little Vermilion Lake 
(92° 6' W, 50° 2.7' N). These rocks have 
been described in some detail by F. J. Pet- 
tijohn (Bull. Geol. Soc. Am., 45, 479, 1934), 
who suggested (personal communication) 
that they represented a very ancient ter- 

A K feldspar-rich fraction was run for 
samples 1 and 2, and a plagioclase-rich 
fraction for sample 3. The analytical data 
and apparent ages are shown in table 4. 
The normal strontium of sample 3, cor- 
rected for the decay of rubidium, was used 
for the normal strontium in the age cal- 
culations (87/88 = 0.0835). From the re- 
sults it can be seen that the apparent age of 
these boulders is no greater than that of 
the other granites shown in table 3 in 
spite of their presumed greater geological 


4. A 


Data and 

Ages of 


Sr 87 




Sr Tota 

1 ■?' 
' million 




















age. It is not clear whether these dates 
represent the age of the source material 
or some later period of alteration. The 
possibility of later alteration is indicated 
by the considerable sericitization of the 
plagioclase feldspars and partial sericitiza- 
tion of the potassium feldspars. It had 
been hoped that these measurements 
would narrow the gap between the oldest 
dated rocks and the presumed age of the 
earth. Obviously, further studies are neces- 
sary in order to find out whether there is 
any record of this interval in terrestrial 

ment on the Levack norite about 50 miles 
to the northeast. Other data obtained at 
Massachusetts Institute of Technology on 
the ages of rocks around the north range 
of the Sudbury nickel irruptive are in 
agreement with this figure. Since the in- 
terpretation of the data of table 2 is still 
obscure, it would be premature to attempt 
a definitive interpretation of data from 
deep in the Precambrian. It seems prob- 
able, however, that the time of intrusion 
of the pegmatite and the granite is indi- 
cated by the 1750-million-year figure and 
that the 1300-million-year ages represent a 

TABLE 5. Ages of Minerals from Rocks of the Cutler Batholith, Ontario 


million years 





Pegmatite A 












Schist cut by pegmatite A 




Pegmatite B 




Granite B cut by pegmatite B 







Granite C 







Granite D 




Levack, Ont., biotite norite about 60 

miles to the northeast 




Zircon from granite B gave the following ages: 





p b 207 


Pb 207 -Pb 206 


Th 232 -Pb 2n8 


Farther to the south, younger orogenic 
belts are represented. Table 5 is a compila- 
tion of our current data on the Cutler 
batholith on the north shore of Lake 
Huron about midway between Sudbury, 
Ontario, and Sault Ste. Marie. Compari- 
son of table 5 with table 2 shows the simi- 
larity of these two sets of data. In both, 
the pegmatite feldspars and muscovites 
have old Rb-Sr ages, and the biotite ages 
and the pegmatite K-A ages are consist- 
ently younger. Some support for the sig- 
nificance of a 1750-million-year date in this 
area can be obtained from the measure- 

subsequent event, or the end of a long 
period of elevated temperatures. The loca- 
tions of the Ontario samples are shown in 
figure 5. 

Michigan. In Year Book 57, data were 
given on the age patterns from the meta- 
morphic zones of a single geologic forma- 
tion. The K-A and Rb-Sr ages obtained 
from micas found in the various meta- 
morphic zones of the Michigamme slate 
in Iron County, Michigan, were shown to 
be in agreement with each other and with 
the age of mica in the granite thought to 
have been intruded at the time of meta- 


morphism. Before these measurements 
were made, it had been believed that the 
intrusion of the granite, the emplacement 
of the pegmatites in near-by Dickinson 
County, and the metamorphism of the 
sedimentary rocks in both counties were 
all contemporaneous. The age measure- 
ments indicate that at least two times of 
mineral formation must be postulated. 

phism. Additional rocks from Dickinson 
County were collected to determine the 
extent of the later metamorphism. The 
locations of the samples that have been 
analyzed are shown in figure 6. The ages 
obtained on various minerals at these loca- 
tions are given in table 6. Except for the 
sample from Republic, all the locations lie 
within a square less than 40 miles on a 

Fig. 5. Map showing locations of samples from Ontario. 

Whether the granite and the pegmatites 
are contemporaneous or not cannot be re- 
solved with the present measurements, but 
the data indicate that the latest time of 
metamorphism of the Michigamme slate 
is significantly later than the time of in- 
trusion of the pegmatites in Dickinson 

This year's work has extended the meas- 
urements to other metamorphic zones of 
the Michigamme slate and to granites 
thought to be related to the metamor- 

side. The geologic formation from which 
each sample was taken is indicated by the 
letters at the site, and the geologic section 
adapted from the extensive work of the 
Geological Survey in the area is shown in 
the key at the side. 

All the ages west and south of Iron 
Mountain were found to be 1400 million 
years or less, no matter what the relative 
ages of the formation were judged to be 
by stratigraphic assignment. This may be 
interpreted simply as an indication of the 


Fig. 6. Localities sampled in upper Michigan. 


effect of the 1400-million-year metamor- 
phism on all the rocks in the area. The 
fact that the total rock analysis of the chlo- 
rite zone, 2, gave similar ages means that 
the Michigamme slate was deposited at a 
time close to the metamorphism or else 
that even the low-grade metamorphism 

TABLE 6. Ages Measured in Upper Michigan 
and Northeast Wisconsin 



Age, millior 

i years 





1430 ± 200 


Total rock 

1300 ± 130 





































(a) Pegmatite, 




(b) Gneiss,* 


































* Zircon from this gneiss 

gave the following 


fJ238_p[- ) 206 


fJ235_p[ 3 207 


p b 207_p b 206 


either introduced most of the rubidium 
and potassium to the slate or removed 
radiogenic Sr 87 and A 40 , thereby erasing 
the previous history of the rock. The low 
biotite age of the oldest granite (8) in this 
area indicates the very real possibility of 
a metamorphism later than that surmised 
from the age of the metamorphism of the 

In the area north of Iron Mountain 
along the line east of Felch (Felch trough) 
a much more complex pattern of ages has 
emerged. The geologic evidence is that the 
pegmatites and granites intrude all the 
sedimentary rocks in the area and that 
these in turn overlie the gray gneiss and 
the older granite gneiss. The sedimentary 
rocks have also been metamorphosed pre- 
sumably at the time of emplacement of the 
pegmatites and granites. The age pattern 
of the minerals from the metamorphosed 
sediments indicates a time of metamor- 
phism 1600 million years or more ago, but 
it is seen that the K-A ages are uniformly 
less than this. This pattern may be in- 
terpreted most simply by the following 
sequence of events : (a) intrusion of gran- 
ites and pegmatites and metamorphism of 
sedimentary rocks in Dickinson County at, 
say, 1800 million years; (b) subsequent 
metamorphism of rocks west and south of 
the Felch trough at 1400 million years. If 
this pattern of ages is really the result of 
two relatively simple periods of mineral 
formation, the ages at site 11 have con- 
siderable significance in establishing the 
stability of minerals in the environment in 
which minerals are formed. The pegma- 
tite clearly intrudes the gneiss. The 2500- 
million-year Pb-Pb age of the zircon from 
the gneiss is evidence that the gneiss is 
really a very ancient rock. The biotite ages 
of the gneiss seem to be due to minerals 
formed at least 300 million years after the 
pegmatite intrusion. Thus the geologic 
event that formed the biotite in the gneiss 
was relatively ineffective in altering the 
age pattern of the pegmatitic muscovite 
and feldspar. At site 15 it is again shown 
that the age pattern of muscovite is some- 
what more resistant to alteration than that 
of biotite in the same environment. It is 
this fact that makes it difficult to state with 
certainty that the granites in Iron County 
are not contemporaneous with the peg- 
matites in Dickinson County, since only 
biotites have been analyzed in the Iron 
County granites. 



Wisconsin. The exposed Precambrian 
of Wisconsin is thought to reflect many 
of the complexities of the buried Precam- 
brian of the central United States. Ac- 
cordingly, an age program on central and 
southern Wisconsin was undertaken, to 
be extended later to the "basement" of the 
Interior Lowlands. At the present time 
only Rb-Sr ages of micas have been meas- 
ured. Conclusions will be drawn only 
when these ages are supplemented by ages 
based on other methods and minerals. 

Wisconsin ages (table 7) are classified 

TABLE 7. Rb-Sr Ages from Wisconsin 
Grouped According to Weidman 

TABLE 7 — Continued 






Basal Group 

Gneiss and schist 

Mill Creek Biotite 1940 

Conants Rapids Biotite 1900 

Nekoosa Biotite 1640 

Port Edwards Biotite 1620 

Black River Falls Biotite 1610 

Neillsville Biotite 1570 

Whiting-Plover Dam Biotite 1540 

Stevens Point Biotite 1320 

Lower Sedimentary Series 
Slate, graywacke, schist, and quartzite 

Merrill Muscovite 1800 

Biotite 1510 

Athens Muscovite 1600 

Biotite 1250 

Rib River Biotite 1390 

Igneous Intrusivt 

' Series * 


Moonstone pegmatite 

quarry west of Wausau 



Wausau, irregular, coarse- 

grained, biotite-rich 

patches grading into sur- 

rounding medium- 

grained basic syenite 




Plover River 



Waterloo pegmatite 



Rapakivi quarry, Waupaca 



Big Falls 








Inclusion, fine-grained, 

almost massive 
Inclusion, coarse- 
grained, schistose 
Wausau, lenticular graphic 
pegmatite with quartz 
core transecting with 
sharp boundary medium- 
grained basic syenite 
listed above 
Pegmatite exposed in 
bank of Eau Claire 
River; grades into sur- 
rounding coarse- 
grained granitic rock 













Problematic Assignment 

Prairie River Dells 

Gray granitic rock, mas- 
sive to foliated; unlike 
pink massive granites 
listed above Biotite 

Massive basic dike cut- 
ting gray rock with 
straight sharp bound- 
ary Biotite 
Trapp River, highly vari- 
able hybrid; where rela- 
tively unaltered, older 
phase similar to slates of 
Lower Sedimentary 
Series Biotite 




* Texture granitic except where otherwise spec- 
ified. No metamorphic foliation. 

according to Weidman's ( Wis. Geol. Surv., 
Bull. 16, 1907) scheme, which is based on 
the latest and most complete regional 
mapping. Specimens not located on Weid- 
man's map are unambiguously assigned to 
his groups, whereas the assignment of 
three specimens from his map area is un- 

The Basal Group yields only metamor- 
phic ages which at most permit the con- 
clusion that these rocks originally crystal- 


lized more than 1940 million years ago 
and were later subjected to one or more 
age-modifying events. Only the Stevens 
Point biotite gave an apparent age less than 
that of the majority of massive igneous 
rocks measured. 

Although he did not observe them in 
contact, Weidman thought that the Lower 
Sedimentary Series is younger than the 
Basal Group. He based his belief on the 
higher degree of metamorphism of the 
Basal Group and other petrologic differ- 

Other U. S. Precambrian Age 

Oklahoma and Missouri. The results of 
the continuation of the work on igneous 
rocks in southern Missouri and the Ar- 
buckle and Wichita Mountains of Okla- 
homa are shown in table 8. The age of 
the Missouri and Arbuckle rocks is ap- 
parently about 1400 million years, roughly 
the same as the ages previously found in 
the exposed Precambrian areas of the 

TABLE 8. Mineral Ages of Rocks from Oklahoma and Missouri 



Age, million years 





p b 207 

Th 232 



p b 206 

p b 207 

p b 206 

p b 208 


Graniteville, Mo., granite 





Fredericktown, Mo., pegma- 





Decaturville, Mo., pegmatite 





Tichomingo, Arbuckle Mts., 
Okla., granite 










Wichita Mts., Okla., Lugert 





Meres, Okla., Wichita Mts., 




Wichita Mts., Okla., pegma- 

tite in Quanah granite 






ences, and on a somewhat speculative re- 
gional structural interpretation. Where the 
two rock groups were observed in the pres- 
ent study, their petrologic differences were 
amply confirmed. Both groups yield meta- 
morphic ages whose ranges overlap so ex- 
tensively that the ages cannot be used as 
support for a considerable difference in 
time of original metamorphism of the 
Lower Sedimentary Series and the Basal 

Weidman lumped all massive igneous 
rocks into his Igneous Intrusive Series. 
Despite possible evidence among the ages 
for subdivision of this series, conclusions 
are perhaps better deferred for the mo- 

Rocky Mountains. The Wichita rocks in- 
dicate much younger ages, and it is doubt- 
ful that they are Precambrian at all. The 
location of the samples is indicated in 
figure 7. 

Southern California. In collaboration 
with G. J. Wasserburg and L. Wright, 









St. Louis J 1 

3* 1 \ .^ 

* *2 V J 

Springfield J._ 


' j Tulsa, 

Q Oklahoma 
! 5.. 7 City 

S -v£ n . 4 

"*"V — ■ •-"■ 


Little • 
Rock / 


Fig. 7. Locations of samples from Missouri 
and Oklahoma. 



measurements are being made on Pre- 
cambrian rocks from Death Valley, Cali- 
fornia. The results are shown in table 9. 
The concordant data from the Silver Lady 
pegmatite indicate an age of about 1750 
million years for this pegmatite. The 
markedly lower age of the fine-grained 
muscovite in the rock which it intrudes is 
another example of the effects reported in 
previous sections that make the resolution 
of close Precambrian geological events very 
difficult. The ages of the other gneiss are 
clearly discordant, and should not be in- 
terpreted as representing any significant 

to have the same age — about 1750 million 
years. Earlier work here and elsewhere 
indicated the same age in eastern Sweden 
(Varutrask), and measurements made by 
Gerling in Leningrad extend this age as 
far east as the White Sea. This region of 
1750-million-year-old rocks has no known 
equal either in area or in consistency of 
the age data. In Karelia the rocks of the 
Karelidic orogenic belt rest on an older 
basement, and it was expected that this 
older age would be reflected by age meas- 
urements. These measurements, as given 
in table 10, however, show that, like the 

TABLE 9. Mineral Ages from Death Valley, California 


million years 





Silver Lady pegmatite, west side of 

Death Valley 

Muscovite 1 



Muscovite 2 





Schist cut by Silver Lady pegmatite 




Granite gneiss, near Shoshone, east 

side of 

Death Valley 




Monarch Canyon pegmatite 



dates. The Monarch Canyon pegmatite 
was believed on dubious grounds to be 
Precambrian, but these measurements 
show that it probably is not. The meas- 
urements are interesting in that, in spite 
of their discordance, they indicate the pres- 
ence of relatively early Precambrian rocks 
near the western margin of the continent. 

Age Measurements on Roc\s from Other 

Finland. In collaboration with J. A. O. 
Kouvo, of the Outokumpu Co., Outo- 
kumpu, Finland, and Paul Gast, of the 
University of Minnesota, a study is being 
made of the ages of the metamorphic and 
igneous rocks of the Karelian basement 
complex. In previous work of Kouvo and 
Gast it had been found that the two rec- 
ognized orogenic belts in Finland, the 
Karelidic and the Svecofennidic, appeared 

Baltimore gneiss, the biotite in these base- 
ment gneisses indicates the age of the 
younger orogeny, while the zircon and 
the feldspar probably reflect the older age 
of the gneiss. About 100 miles farther to 
the east Gerling has found muscovites giv- 
ing K-A ages of 2600 million years, and so 
it may be expected that continuation of 
this work will lead to a more definitive 
establishment of a 2600-million-year oro- 
genic period in this region, however ob- 
scured by the pervasive 1750-million-year 
event. The locations of the samples are 
shown on figure 8. 

Saudi Arabia. Another group of rocks 
whose age pattern indicates the possibility 
of broad regional metamorphism is shown 
in table 11. These measurements were 
made on a suite of samples collected from 
the Arabian shield by Glenn F. Brown, 
of the U. S. Geological Survey. Mr. Brown 


TABLE 10. M 

neral Ages 

of Rocks from Fin 

and and Vicinity 



Age, mill 

on years 





Pb 207 





p b 207 

p b 206 

Pb 208 


Matasvaara, basement gneiss 





Koli, basement gneiss 







Heinavaara, gneiss dome 





Sotkuma, gneiss dome 






Pieni Neulamaki, gneiss dome 





Inari, quartz dionite 





Pielavesi, pyroxene granite 





Karelsky Mine, Karelian 




* Zircon age from J. A. O. Kouvo, Bull, comtn. geol. Finlande, 182 (1958). 
f Sample obtained from E. K. Gerling for interlaboratory comparison. 

has been mapping this shield as part of 
the Point IV program during the past ten 
years. The interest in these rocks stems 
not from a presumed metamorphism but 
rather from the broader concern in the 

distribution of ages found in different 
shield areas throughout the world. 

Figure 9 shows the location of the sam- 
ples in Saudi Arabia. First it is seen that 
the Rb-Sr and K-A ages are generally 
discordant. Second, the Rb-Sr ages show a 
fair correlation with their distance east 
from the coast, the age decreasing with 
distance. Finally, the age pattern is quite 
similar to that found in the Appalachians 
starting in the New York Highlands and 
ending in the Philadelphia area (table 1). 
Since the younger rocks have not yet been 
measured (or located) the analogy with 
the Appalachian rocks is hardly on a 
sound base, but until more is known about 
the effect of a later orogeny on an existing 
age pattern of a rock this speculation is 

TABLE 11. Ages of Micas from Rocks from 
the Arabian Shield 




Age, million years 


Fig. 8. Map showing locations of samples 
from Finland and Karelian S.S.R. 














Fig. 9. Locations of samples from Saudi Arabia. 

Western Australia. Table 12 shows ages 
of samples from Western Australia meas- 
ured at Carnegie. In Year Book 55 the 
analyses of Jeffery showed the occurrence 
in that area of rocks 2700 million years old. 
These measurements indicate the pres- 
ence of younger rocks, demonstrating that 
areas with complex patterns of mineral 
ages will be available for study in West- 
ern Australia. 

Venezuelan Andes. Rb-Sr ages of five 
minerals and a whole rock from the Vene- 
zuelan Andes have been measured. The 
specimens were collected by Dr. Reginald 
Shagam, of the Venezuelan Ministry of 

TABLE 12. Mineral Ages from Western 








Yinnietharra, main 

open cut 





Morrisey Hill 




Fraser Range 




Hines Hill 




Mines and Hydrocarbons, from an area he 
is mapping. On the basis of geologic evi- 
dence, the gneisses, schists, and granitic 
rocks of the Sierra Nevada formation are 
the oldest known rocks. The Mucuchachi 
phyllite unconformably overlies the Sierra 
Nevada formation. Middle Devonian fos- 
sils were found recently in the Mucucha- 
chi phyllite (personal communication, 
G. R. Pierce, Creole Petroleum Corpora- 
tion, Caracas, Venezuela). From the data 
obtained so far (table 13) it can only be 

TABLE 13. Venezuelan Andes, Rb-Sr Ages 

Rock Mineral 


million years 

Sierra Nevada formation 

Schist Biotite 




Gneiss Biotite 



schist Biotite 




Mucuchachi phyllite 

Whole rock, carbo- 
naceous phyllite 

450 ±95 

said that the Sierra Nevada formation was 
metamorphosed prior to 445 million years 
ago. The latest event it records is younger 
than 235 million years. 



Donald H. Lindsley i 

Basalts and other basic lavas are gen- 
erally considered the most satisfactory 
igneous rocks for determining past posi- 
tions of the earth's magnetic poles. There 
is ample evidence that lavas extruded in the 
last several hundred years acquire a mag- 
netic polarization in the direction of the 
ambient field at the time of cooling. Most 
workers have been content to assume that 
prehistoric lava flows likewise acquired a 
polarization parallel to the geomagnetic 
field, and, further, that they retain the 

4 Carnegie Institution Fellow; from Johns 
Hopkins University; part time, 1957-1959. 


original polarization over considerable 
periods of time. Ten years ago Graham 
established criteria by which field tests of 
magnetic stability of rocks might be made, 
and more recently Nagata and his co- 
workers proposed several laboratory tests 
that should be satisfied before the polar- 
izations of lavas are used for paleomag- 
netic studies. 

Although field and laboratory tests are 
useful as aids in the interpretation of ob- 
served magnetic polarizations in rocks, 
they add little to our understanding of 
why certain flows are magnetically stable 
and others not. Knowledge of any prop- 
erties of lavas (and their component min- 
erals) which are related to magnetic sta- 
bility or instability would be of great 
assistance in choosing rock units for paleo- 
magnetic studies. Ideally, it would be 
helpful to know from field observation 
whether a given formation is likely to be 
suitable. Petrographic studies might yield 
valuable clues, and conceivably the bulk 
chemical composition of the lava might 
be related to stability. Frequently field, 
petrographic, and chemical descriptions 
are available in published material, but 
sometimes it might be faster and cheaper 
to make preliminary geologic studies be- 
fore undertaking a complete paleomag- 
netic investigation. 

With these considerations in mind, we 
decided to make a study of a well exposed 
sequence of basic lavas, of relatively young 
geologic age, whose structural history was 
sufficiently simple to permit a reasonably 
accurate interpretation of the stresses to 
which the rocks had been subjected. The 
Spray Quadrangle of north central Oregon 
was chosen for the investigation. In it 
are exposed basaltic andesites of the Clarno 
formation (late Eocene to early Oligocene 
in age), an olivine basalt flow and feeder 
dike in the John Day formation (late 
Oligocene or early Miocene), and about 
twenty flows, well exposed, of the Mio- 
cene Columbia River basalt. The quad- 
rangle was mapped geologically on the 

scale of 1 inch to the mile, and strati- 
graphic sections of the Columbia River 
basalt were measured in several deep can- 
yons in the southern part of the area. Of 
193 oriented specimens collected for mag- 
netic and petrographic studies, 170 were 
taken from the Columbia River basalt; 
these samples are accurately located, both 
in relation to the top and bottom of the 
flow in which they are found and as to 
their position within the entire sequence. 
The sample number was painted on the 
outcrop so that the exact location could 
be found if the outcrop had to be revisited. 
Magnetic properties of samples from dif- 
ferent parts of one flow as well as those 
from different flows can therefore be com- 

The magnetic orientation and intensity 
of the samples were measured in the De- 
partment's portable spinner magnetometer, 
the directions being plotted on the Schmidt 
equal-area net. Twelve samples from six 
flows of the Clarno formation showed 
widely divergent polarizations. Correc- 
tion for tilt (geologic dip) of the flows in- 
creased this scatter, an indication of mag- 
netic instability, as shown by Graham. 
Plots from ten samples of the John Day 
olivine basalt were distinctly to the south 
and east of the present geomagnetic field. 
Correction for tilt did not reduce the 

Columbia River basalt. Field and pet- 
rographic criteria permit a division of the 
Columbia River basalt in the Spray area 
into a lower "greasy phase" and an upper 
"normal phase." There is evidence that 
the flows of the lower unit were unusually 
rich in volatile components. In flow 2 of 
the greasy phase, escaping volatiles altered 
the margins of the fissures (joints) along 
which they escaped, during which process 
most of the magnetite in the joint margins 
was oxidized to hematite. From petro- 
graphic studies, it seems likely that the 
flows of the greasy phase are richer in 
titanium and possibly in magnesium than 
the normal Columbia River basalt. Chemi- 


cal analyses are being made to check these 

When stability tests (tilt correction) are 
applied to flows of the greasy phase, an 
increase in the scatter of the plots indi- 
cates at least partial magnetic instability. 
Where such tests have been made in the 
upper unit they show magnetic stability. 
To check the results, samples were de- 
magnetized in alternating fields up to 130 
oersteds. The instrument used for demag- 
netization, kindly supplied by the U. S. 
Geological Survey, rotates the sample about 
three perpendicular axes while the field is 
slowly reduced from a chosen maximum 
(Hmax) to zero. The intensity of mag- 
netization of the sample is plotted as a 
function of Hmax. It was thought that the 
demagnetization would destroy any iso- 
thermal or viscous magnetization acquired 
by the basalts after cooling, and would 
leave only thermoremanent magnetization. 
In general, the demagnetization curves for 
samples from the "normal phase" of the 
Columbia River basalt showed a decrease 
in intensity with increasing Hmax, and the 
directions of polarization showed little 
change. Results for all but the upper 
two flows of the greasy phase are more 
complex. The demagnetization curves of 
most samples from the lower unit show 
an initial decrease in intensity with increas- 
ing field, but commonly exhibit an irregu- 
lar increase in intensity as Hmax reaches 
60 to 100 oersteds. 

Magnetic measurements were made on 
samples from six flows of the greasy phase 
and eight flows of the normal phase of the 
Columbia River basalt. The means of the 
plots from each of the fourteen flows are 
different from the direction of the present 
geomagnetic field in the Spray area. The 
mean plots are centered around a direc- 
tion corresponding to an axial dipole. 

Campbell and Runcorn (/. Geophys. 
Research, 1956) have reported alternating 
units of normal and reversed flows in the 
Columbia River basalt. Runcorn con- 
sidered these results as indicating at least 
three reversals in the direction of the geo- 

magnetic field while the basalts were be- 
ing extruded in Miocene time. In only 
one flow (flow 2) of fourteen measured 
from the Columbia River basalt in the 
Spray area was there found an appreciable 
number of samples with reversed polariza- 
tion, and they are outnumbered by normal 
(north-seeking downward) samples from 
the same flow — twelve of thirty-two sam- 
ples have reversed polarization. The re- 
versed samples come mainly from three 
parts of the flow: the base, the top, and 
the altered joint margins described above. 
Because of this fact, I had originally 
reached a tentative conclusion that the top 
and bottom samples (which show great 
scatter) had been moved mechanically 
after cooling, and that the joint margins 
(the relatively small scatter of which is 
further reduced by tilt correction, and 
which contain both magnetite and hema- 
tite) had undergone a self-reversal proc- 
ess such as one of those suggested by Neel. 
The demagnetization curves for samples 
from flow 2 suggest a more likely explana- 
tion. For samples with reversed polariza- 
tions, including those from the altered 
joint margins, the moment increases with 
increasing Hmax. Samples with normal 
polarization show a decrease in intensity 
up to values of Hmax = 60 to 80 oersteds. 
Still higher field intensities produce an in- 
crease in intensity of the sample, which is 
accompanied by a reversal in the polariza- 
tion. The most likely explanation of these 
phenomena is that the entire flow origi- 
nally possessed reversed polarization and 
that the magnetite in most of the flow was 
magnetically unstable, whereas the hema- 
tite in the altered joint margins was stable, 
and the fine-grained magnetite at the top 
and bottom of the flow was at least partly 

The preceding interpretation does not 
preclude the possibility that flow 2 was 
extruded during a period of normal geo- 
magnetic field and that the entire flow has 
undergone a subsequent self -reversal. This 
flow, however, which attains a maximum 


thickness of about 280 feet, shows a wide 
variation in the grain size and composi- 
tion of its magnetic constituents. Any self- 
reversing process that might have affected 
this entire flow must be considerably more 
subtle than any that have previously been 
proposed. Therefore it seems likely that 
the polarity of flow 2 reflects cooling dur- 
ing a period of reverse geomagnetic field. 
Parts of this project yet to be completed 
include statistical analysis of the magnetic 

data, detailed petrographic study of the 
samples, determination of Curie tempera- 
tures of about ten representative samples, 
chemical analyses of fourteen samples, and 
X-ray examination of the magnetic min- 

For assistance on this project, financial 
and otherwise, I wish to thank the Na- 
tional Science Foundation and the U. S. 
Geological Survey, as well as the Depart- 


S. E. Forbush 


As was stated in last year's report, an 
investigation of the equatorial electrojet 
was begun as part of the United States 
program for the International Geophysi- 
cal Year (IGY) with funds granted by 
the IGY Panel on Geomagnetism. It is 
planned to continue the observational pro- 
gram under the International Geophysical 
Cooperation Year (IGC) for 1959. Eco- 
nomies effected during the IGY (July 1957 
through December 1958) permit this ex- 
tension without additional funds. 

The four Askania variographs have op- 
erated without major difficulties during the 
report year. Some delay was encountered 
in effecting reliable operation of the new 
magnetic observatory at Arequipa, Peru. 
This observatory was established by the 
Instituto Geofisico de Huancayo coopera- 
tively with the Universidad Nacional de 
San Agustin, Arequipa. Despite this de- 
lay there will be a period of roughly a year 
(1959) during which it should be operat- 
ing concurrently with the Askania vario- 
graphs. In addition to providing useful 
additional data for electrojet studies this 
concurrency will greatly enhance the value 
of data from the Arequipa Observatory 
after operation of the Askania network 
ceases at the end of 1959. 

One graphical computer-scaling device 
was completed for the Askania magneto- 
grams and one for the Huancayo mag- 
netograms. These devices permit direct 

reading of final values in less time than 
was previously required for conversion to 
absolute values, and the addition of base- 
line values is saved. 

Preliminary studies indicate that the 
magnetic-disturbance diurnal variation, Sd, 
sometimes, but not always, exhibits electro- 
jet effects; that is, the Sd variation is some- 
times augmented under the electrojet band 
in about the same way as the quiet-day 
diurnal variation, S q , always is. On the 
other hand, there are occasions when Sd 
variations appear only at stations too far 
from the electrojet region to be influenced 
by it — an indication that the current sys- 
tem for Sd does not always extend from 
middle latitudes all the way to the equator. 

Determining whether the semidiurnal 
lunar variation, L, in horizontal intensity 
is, like S Q , enhanced under the electrojet 
band is of considerable interest in con- 
nection with the question whether the cur- 
rent system for L flows at the same height 
as that for Sq. Owing to the variability in 
amplitude and phase of S q at Huancayo, 
for example, extensive data are required 
for the determination of the semidiurnal 
lunar variation, although that is quite large 
at Huancayo in certain seasons of the year. 
At Yauca, south of Huancayo, and at 
Chimbote, north of Huancayo, the diurnal 
variations in horizontal intensity, H, are 
similar to but smaller than those at Huan- 
cayo. On the other hand, the diurnal 
variations in vertical intensity, Z, at Yauca 


and Chimbote are opposite in phase. If 
the diurnal variation in Z at Yauca is sub- 
tracted from that at Chimbote the dif- 
ference, say Saz, is similar and comparable 
in magnitude to the diurnal variation Sh 
in H at Huancayo. The daily values of 
Saz and Sh may vary by a factor of 2 on 
account of the variability of S q . The high 
correlation between daily values of Saz 
and Sh permits, in effect, the elimination 
of the variability of S q . Thus, if the lunar 
variation in Saz were identical to that in 
Sh, and there were no other disturbing in- 
fluences, Sh and Saz would be perfectly 
correlated. On the other hand, if, for ex- 
ample, there were a lunar effect in Sh but 
none in Saz, the departures from perfect 
correlation between Saz and Sh would de- 
termine the lunar variation in Sh from 
data freed, in effect, of the large variability 
of S Q . The lunar variation in H at Huan- 
cayo is well determined from Bartels' 
analysis of extensive data. Making use of 
this and the correlation between Saz and 
Sh, the lunar effect in Saz can be evalu- 
ated, and this, in turn, would determine 
the extent of electrojet effects on the semi- 
diurnal lunar variation. From an inade- 
quate sample of data, preliminary deter- 
minations in this manner did not give 
definitive results. 


World-wide changes in cosmic-ray in- 
tensity and in magnetic horizontal in- 
tensity. From the Department's cosmic- 
ray program it has been found that some, 
but not all, magnetic storms are accom- 
panied by world-wide decreases in cos- 
mic-ray intensity. During the main phase 
and recovery of individual magnetic 
storms the changes in daily means of cos- 
mic-ray intensity at Huancayo, Chelten- 
ham, Christchurch, or Godhavn were 
found to be fairly well correlated with 
changes in the daily mean magnetic-storm 
field as determined by daily means of mag- 
netic horizontal intensity at Huancayo. 
The ratio between these changes, however, 

differs in different storms, and, indeed, in 
some no detectable cosmic-ray changes are 
observed. Nevertheless it is of interest to 
determine whether bihourly means of cos- 
mic-ray intensity follow closely the bi- 
hourly means of the magnetic-storm field. 

Heretofore no reliable measures of the 
world-wide component of the magnetic- 
storm field have been available for periods 
of less than a day, owing to the diffi- 
culty of separating the world-wide com- 
ponent of the storm field from the or- 
dinary but variable diurnal variation, S q , 
and the often large and variable disturb- 
ance variations, Sd, superposed upon it. 
Using only night-time values of horizontal 
magnetic intensity, in which S Q and Sn 
effects are small, from four equatorial 
magnetic observatories (Huancayo, Wa- 
theroo, Apia, and Elizabethville) sepa- 
rated about 6 hours in longitude, Walter 
Kertz has recently derived and published 
new 3-hour mean indices, U , for the field 
strength of the equatorial ring current, 
for the period January 1939 to December 
1945. A comparison of these U measures 
with the bihourly means of cosmic-ray in- 
tensity (from Huancayo) during a few 
storms between January 1939 and Decem- 
ber 1945 indicates that the bihourly cos- 
mic-ray changes are not closely correlated 
with the 3-hour U values. The implica- 
tion is that the cosmic-ray changes are not 
directly due to the magnetic-storm field. 
This lack of close correspondence is fur- 
ther evidence that the decrease in cosmic- 
ray intensity during magnetic storms may 
arise from the earth's becoming enveloped 
in large plasma clouds from the sun dur- 
ing the storms. These clouds contain 
"frozen-in" magnetic fields and a "de- 
ficiency" of cosmic rays due to their exclu- 
sion by solar magnetic fields. 

Sudden decrease of cosmic-ray intensity 
at the equator and at high latitude. In the 
period from September 1956 to December 
1957, seven large, sudden decreases and re- 
coveries in cosmic-ray intensity were com- 
pared by means of data from the ioniza- 
tion chamber at Huancayo and from a 


neutron monitor at Uppsala, Sweden. On 
the basis of percentage changes, the de- 
creases and the recoveries during these 
storms averaged about 2 l / 2 times larger at 
Uppsala than at Huancayo. There were 
significant deviations from this ratio in 
individual storms, indicating that the en- 
ergy spectrum of excluded primary par- 
ticles may differ in different storms. The 
analysis has been extended, using IGY 
neutron data from Ottawa, to include 1958. 
The ratio of the yearly average of the de- 
creases at Uppsala (or Ottawa) to that at 
Huancayo was found to be the same as 
for the yearly average of the subsequent 
recoveries after minimum. The ratio for 
1956, however, is found to be about 12 per 
cent greater than for 1957 and 1958. The 
difference probably arises from the fact 
that the neutron intensity was, owing to 
the large 11-year variation, about 12 per 
cent greater at Uppsala (and Ottawa) in 
1956 than in 1957 and 1958, and from the 
fact that the higher neutron intensity at 
Uppsala (or Ottawa) in 1956 was due to 
low-energy primaries that are excluded by 
the earth's field at Huancayo. Evidently, 
therefore, these additional low-energy pri- 
maries at high latitudes are excluded dur- 
ing magnetic storms. 

Cosmic-ray diurnal variation. The long 
series of ionization-chamber data from 
Huancayo, Cheltenham (or Fredericks- 
burg), and Christchurch have been har- 
monically analyzed to obtain yearly means 
of the 24-hourly and 12-hourly waves of 
cosmic-ray intensity corrected for pres- 
sure. The analyses were made principally 
to determine in particular the nature of 
the apparent secular change in the 24-hour 
wave which is similar at all three stations, 
except at Huancayo for 1946 to 1958. This 
discrepancy was found to arise from a sys- 
tematic error in the application of correc- 
tions for barometric pressure at the center 
of each 2-hour interval. After 1946, means 
of instantaneous values of pressure at the 
hours n and (n + l) were inadvertently 
used to correct the bihourly mean of ion- 
ization over the interval (« — 1) to (n + l). 

Thus the pressure correction applied was 
appropriate for a bihourly mean of ioniza- 
tion centered l / 2 hour too late. Fortunately 
the resulting error in the published 2-hour 
means, of ionization corrected for pres- 
sure, is negligibly small. It is of conse- 
quence only in its effect on yearly (or 
monthly) averages of the diurnal varia- 
tion. This defect has now been thoroughly 
corrected by means of harmonic coef- 
ficients for the yearly means of pressure 
and ionization. Correction has been made 
as well for timing errors on the baro- 
grams, of which the extremes (averages 
for a year) were 4-22 and —5 minutes. 
It has been found that at Huancayo the 
residual small 12-hourly wave in cosmic- 
ray intensity possibly results from a small 
amount of friction in the pen-and-ink 
recording barograph used at Huancayo; 
corrections are being investigated for this 

Old cosmic-ray program. Compton- 
Bennett meters were satisfactorily operated 
throughout the report year at Godhavn 
(Greenland), Climax (Colorado, U. S.), 
Ciudad Universitaria (Mexico, D. F.), 
Huancayo (Peru), Christchurch (New 
Zealand), and Fredericksburg (Virginia, 

The scaling and reduction of records in- 
cluding the tabulation of bihourly means 
of ionization corrected for barometric pres- 
sure have been completed for Huancayo 
and Fredericksburg through April 1959. 
From the records scaled at Christchurch 
the reduction of daily means has been ef- 
fected through April 1959. To cooperate 
in the United States program for cosmic- 
ray research in the International Geo- 
physical Year, tabulations of corrected bi- 
hourly means of cosmic-ray intensity have 
been forwarded to the four IGY World 
Data Centers. These include IGY data 
for Huancayo, Ciudad Universitaria, Fred- 
ericksburg, and Godhavn. Data from 
Christchurch are supplied by the Christ- 
church Magnetic Observatory to the IGY 
World Data Centers. In addition, in ac- 


cord with recommendations of the U. S. 
IGY Technical Panel for Cosmic Rays, 
data for Godhavn have been reduced for 
the years 1954, 1955, and the first half of 

Large ionization chamber. The large 
cosmic-ray ionization chamber at Derwood 
was operated until May 1959, at which 
time it was dismantled. No solar-flare ef- 
fects have been observed since February 

Cooperation in operation of cosmic-ray 
meters. The successful operation of 
Compton-Bennett cosmic-ray meters over 
a long period at so many stations has been 
possible only through the wholehearted 
and unselfish cooperation of several or- 
ganizations and individuals. We wish to 
express our appreciation to the following 

organizations for the operation and main- 
tenance of cosmic-ray meters: the Danish 
Meteorological Institute and the staff of 
its Godhavn Magnetic Observatory at 
Godhavn, Greenland; the U. S. Coast 
and Geodetic Survey and the staff of its 
magnetic observatory at Fredericksburg, 
Virginia; the High Altitude Observatory 
of the University of Colorado and its staff 
at Climax, Colorado; the Instituto Na- 
cional de la Investigacion Cientifica and 
the Universidad de Mexico, Mexico, D.F.; 
the Government of Peru and the staff of 
its Instituto Geofisico de Huancayo for 
making available the Compton-Bennett 
records from Huancayo; and the Depart- 
ment of Scientific and Industrial Research 
and the staff of its Magnetic Observatory 
at Christchurch, New Zealand. 



N. P. Heydenburg, G. M. Temtner, and 
J. A. Weinman 5 

Polarized Ion Source Development 

Molecular beam apparatus (APIS). As 
was mentioned in our last annual report, 
we became actively interested in opening 
new horizons for our electrostatic accel- 
erator by installing a source of at least 50 
per cent polarized particles in its high- 
voltage terminal; we have discussed pre- 
viously the unique features of our machine 
that make it possible to contemplate even 
bulky and unrefined versions of various 
schemes for the production of polarized 
particles by atomic means. With this ob- 
ject in view we made a number of modifi- 
cations in the terminal in order to provide 
additional space and to raise the electric 
power available locally to nearly 7 kw. 

Meanwhile, in continuation of the joint 
effort with Vernon W. Hughes and 
Charles W. Drake, of Yale University, on 
a molecular beam type of apparatus, one 
of us (J. A. W.) spent several months at 

5 Carnegie Institution Fellow; from University 
of Wisconsin. 

Yale to try to perfect a device for the ul- 
timate ionization of the polarized hydro- 
gen atoms. This step turns out to be both 
one of the most important and the most 
troublesome, since the efficiency of known 
and tried ionizers has been of the order of 
10~ 3 , a factor that this scheme can ill af- 
ford. The ionizer we perfected works in 
part on the principle of the Phillips ion 
gage; i.e., electrons are emitted from a 
filament, attracted to a positive electrode, 
but confined by a magnetic field in such a 
way that they oscillate along the field lines, 
thus having many passes at the neutral 
atoms traversing the region. The device 
has yet to be tried under actual operating 
conditions. The molecular beam apparatus, 
built at Yale entirely of stainless steel, with 
metal gaskets, differentially pumped in 
three stages by mercury diffusion pumps 
to minimize background hydrogen pres- 
sure at the ionizer (the presence of which 
always dilutes any polarized proton beam), 
is essentially completed and is about to be 
tried for a beam. Since the beam axis is 
horizontal, an adapter is needed to bend 
the beam into the vertical direction re- 
quired by our accelerator; this adapter, 


using electrostatic bending plates, is ready 
and has been checked with an ordinary 
ion source beam. 

An ingenious suggestion made recently 
profoundly simplifies the testing procedure 
of the apparatus. It turns out, as a con- 
sequence of essentially geometrical con- 
siderations, that the strong and well known 
resonance in the reaction of deuterons upon 
tritons, lying at only 107 kilovolts bom- 
barding energy and having a prolific neu- 
tron yield, will show a strong anisotropy 
in its neutron angular distribution — easily 
a factor of 2 between the angles 0° and 
90° to the beam — provided that the inci- 
dent deuteron beam is polarized (or at 
least aligned) with respect to the beam 
axis. It is of course isotropic for unpolar- 
ized particles, since the resonance has zero 
orbital angular momentum in its incident 
channel. The effect is to be ascribed to the 
fact that the deuteron has spin unity, and 
hence carries with it a higher angular 
"complexity" than a proton of spin Y z , in 
fact sufficient to allow an angular anisot- 
ropy of terms up to sin 2 0. Now it is 
clear that, by adding a small section of 
accelerating tube to the molecular beam 
apparatus, it will be possible to accelerate 
the finally ionized deuterons to the reso- 
nance energy, observe the neutron anisot- 
ropy, and hence verify the existence of 
alignment of the beam; previously, this 
would have had to be carried out in the 
terminal of our generator, with all the 
concomitant difficulties. Hence all pre- 
liminary tests and adjustments will be 
made at Yale where the apparatus now 

Metastable hydrogen-atom beam appara- 
tus. About November of 1958, we became 
aware of a suggestion by L. Madansky 
(Johns Hopkins University) for an alter- 
native approach to the problem of polar- 
ized ion production. After our visit there, 
L. Madansky, G. E. Owen, and we agreed 
upon a collaborative effort in view of 
the facts that their preliminary experi- 
ments on the production of metastable 

atoms looked promising and relatively 
simple compared with APIS and that our 
uniquely suited accelerator would permit 
an early tryout of the method. Madansky 
and Owen measured a number of prelim- 
inary properties of their device, which is 
basically constituted as follows. 

A conventional radiofrequency ion 
source produces a well focused beam of 
protons of about 10 kilovolts energy. The 
beam is passed through a region of gas 
at about 10 -3 mm Hg, which neutralizes 
a good portion of the beam by electron 
pickup (this is the exact inverse of the 
gas "stripper" device we have used in the 
past to great advantage for the production 
of doubly ionized helium beams; see Year 
Book 54). Now it turns out that not only 
are these hydrogen atoms found in their 
ground state (1/), but a fraction (between 
0.1 and 1 per cent) are left in the meta- 
stable 2s state, the state made famous by 
the brilliant experiments of Lamb and 
Retherford about 10 years ago ("Lamb 
shift"). Forgetting for the moment the 
atoms in the ground state, the metastable 
atoms can be polarized rather simply, as 
pointed out in their first paper. The meta- 
stable 2s state splits in a magnetic field 
into two components corresponding to the 
two possible orientations of the electron 
spin; at an appropriate field, one of these 
components can be made to intersect one 
of the components of the near-by 2p state 
which has a very short lifetime for decay 
by Lyman-alpha radiation to the ground 
state (1.6 X10~ 9 second). By applying a 
small electrostatic field to the beam, a 
Stark effect mixing of the two states will 
take place and the 2s state in question can 
be depleted at will ("quenching"). At the 
same time, the other 2s component will 
have a lifetime up to 2000 times longer, so 
that an essentially 100 per cent polarized 
beam of atoms can easily be achieved. Al- 
though the protons are still randomly ori- 
ented with respect to these oriented elec- 
trons (half up, half down), upon leaving 
the H field region a quantum-mechanical 
mixing of the state with antiparallel pro- 



ton and electron spins will occur, so that 
finally there will be 3 protons pointing up 
for every proton pointing down, a polari- 
zation of 50 per cent. We thus have a 50 
per cent potentially polarized proton beam 
(inside metastable atoms) along with an 
unpolarized beam of ground-state hydro- 
gen atoms, exceeding the former in in- 
tensity by a factor of about 100 or more. 
Now since the ionization energy of a 
metastable atom is only 3.4 volts, and the 
ground state requires 13.6 volts, the pos- 
sibility exists of ionizing only the atoms 
of interest, say by means of a strong ultra- 
violet light source. The ionization limit 
corresponds to about 3650 A. We set up a 
high-pressure mercury arc of very small 
dimensions, emitting about 100 watts of 
photon power in the near ultraviolet. Two 
/0.3 parabolic, front-aluminized mirrors 
could transfer almost 50 per cent of the 
emitted power with a magnification of 
unity to the beam location through a Vy- 
cor tube. 

The Madansky-Owen apparatus was 
brought to the Department in February 
of this year for further experimentation. 
We soon discovered a major flaw in the 
scheme. After the neutral beam (both Is 
and 2s) had traversed a distance of some 
50 cm in the vacuum system, a small frac- 
tion thereof became ionized by collisions 
with the residual gas, so that a sizable 
current was measured in the Faraday cup 
even though all initially charged particles 
had been deflected out by a magnet. It 
turns out that the effect was just as could 
be calculated from the well known atomic 
charge -exchange cross sections. This is 
rather disastrous, since the vacuum re- 
quired to bring the charge exchange be- 
low the amount available from photo- 
ionization of metastables (the latter effi- 
ciency is at best 10 -4 ) is about 10" 9 mm 
Hg or less. These conditions, though hav- 
ing been achieved in the laboratory, are 
by no means available in a routine way, 
especially when gas is admitted to a sys- 
tem continuously. 

Trying to salvage what we could from 

these adversities, we proceeded to find out 
whether there was a sufficiently large dif- 
ference in the "stripping" probabilities of 
2s and Is atoms so that the polarized com- 
ponent could be preferentially ionized. By 
applying a 1000-cycle electric field to the 
beam, we can produce a "chopped" beam 
of metastable atoms; i.e., during the part 
of each cycle when the full voltage is ap- 
plied, only Is atoms emerge (the 2s atoms 
being quenched by the electric field), 
whereas the usual mixture prevails when 
no voltage is applied. If the stripping 
cross section for 2s atoms is larger than 
for Is atoms — not an unreasonable hope 
in view of the fact that the radius of 
the former is 4 times that of the latter, 
and hence presents 16 times the cross- 
sectional area — then a modulated compo- 
nent should be present in the current de- 
tected in the Faraday cup after traversing 
a suitable stripping gas region and hence 
could be measured with a phase-sensitive 
detector. Our preliminary experiments do 
not reveal any appreciable difference in 
stripping cross sections, so that this scheme 
cannot be used to advantage at present. 

Another scheme to make use of the 
stripping process for ionization is to have 
a molecular beam device of the type de- 
scribed in the first section, but form the 
ground-state neutral beam from a focused 
beam of very low-energy protons as in the 
present section, thus avoiding the inverse- 
square loss of intensity. The energy would 
have to be a compromise between useful 
values of electron-loss cross sections, maxi- 
mum practical size of gradient Stern- 
Gerlach magnets to produce a spatial sep- 
aration of the beam components, and 
possibility of intense ion beam formation. 
The existence of such a compromise en- 
ergy seems doubtful at best. 

We are carrying out a number of atomic 
physics measurements in the hope of dis- 
covering properties that would be useful 
in the production of polarized ion beams, 
measurements that are of interest in them- 
selves, since these "nuclear" techniques are 



practically unexplored in the physics of 


Because the alterations in the cyclotron, 
mentioned in Year Book 52, and the addi- 
tion of focusing magnets (Year Book 55), 
failed to give the desired increase in ex- 
ternal beam needed for scattering experi- 
ments, tests were made on the fall-off of 
beam intensity within the dee region. It 
was determined that only a small fraction 
of the initial beam ever reached the radius 
corresponding to the position of the beam 
septum and deflector. A check on the 
magnetic field intensity in the deflector 
region indicated that the field fell below 
the proper value for resonance before the 
beam could reach the deflector. In the 
previous alterations the 1-inch filler plates 
on the top and bottom poles had been re- 
moved to make room for 1-inch rim shims. 
The increase in current in the magnet coils 
to compensate for this reduction of iron 
resulted in greater saturation of the iron 
near the edge of the poles, which would 
account for the field fall-off observed. 

To rectify the difficulty, the 1-inch filler 
plates have been replaced. Some iron pre- 
viously removed from the pole faces at 
the outer rim has been restored, and new, 
shallower rim shims have been added. The 
addition of the filler plates and the rim 
shims has reduced the gap between poles 
to the point where new and thinner dees 
have become necessary. New dees and 
a different deflector system will be con- 
structed in the near future. 


E. T. Bolton, R. /. Britten, D. B. Cowie, 

B. J. McCarthy, 6 K. McQuillen, 7 and 

R. B. Roberts 


For the last two years an increasingly 
large share of the attention of the Bio- 
physics Section has been devoted to the 

G Carnegie Institution Fellow; from Oxford 

7 Research Associate, Carnegie Institution of 

Washington; from Cambridge University. 

study of ribonucleoprotein particles found 
in bacteria. This particular aspect of our 
general interest in the synthesis of large 
molecules by growing cells has been em- 
phasized for a number of reasons. Similar 
particles are found in tissues of all kinds, 
including microorganisms, plants, and ani- 
mals. Their concentration is highest when- 
ever the rate of protein synthesis is highest, 
and it now seems that they provide the 
sites for the synthesis of most, if not all, 
protein. Accordingly, they appear to pro- 
vide the templates for protein synthesis 
and are the active machinery responsible 
for the formation of enzymes. It is well 
established, however, that the capability 
to form enzymes is determined by the 
genetic constitution of the cell, i.e., by its 
deoxynucleic acid (DNA). Thus the ribo- 
nucleoprotein particles (ribosomes) must 
act as intermediaries between the DNA 
which carries the information needed to 
specify the synthesis of an enzyme and 
the enzyme which may be formed when 
a certain gene is present. 

It is not known whether the DNA also 
participates directly in the synthesis of an 
enzyme, whether DNA necessarily par- 
ticipates in the synthesis of the ribosomes, 
or whether, once initiated, the ribosomes 
can continue an autocatalytic synthesis 
which is independent of DNA. It does 
seem, however, that further study of the 
ribosomes, of their composition and struc- 
ture, of their synthesis and their variation, 
of their enzymic properties, of their action 
in the synthesis of protein will give an- 
swers to a number of these questions. 

Many aspects of the ribosomes previ- 
ously explored here in a qualitative way 
have been investigated more quantitatively 
during this year. The sedimentation con- 
stants of the various classes of ribosomes 
and their variation with conditions, both 
in the cell and in the preparation of the 
ribosomes, have been measured. The 
methods of analysis have been improved. 
In particular, a method has been devised 
for the separation of the ribosomes of 
different sedimentation constants. This 



method has been a great help in the study 
of the composition of ribosomes, of their 
role in protein synthesis. 

Using very short periods of incorpora- 
tion of radioactive tracers, it has been 
possible to observe nascent protein which 
is associated with the ribosomes for a short 
period before it is released as a finished 
product. Thus the ribosomes of bacteria 
have the same role in protein synthesis as 
had previously been demonstrated for the 
ribosomes of higher organisms, but the 
bacterial ribosomes work roughly 60 times 

The separation of the various classes of 
ribosomes has made it possible to show 
clearly that the synthesis of the ribosomes 
themselves is an intricate process. Newly 
formed nucleic acid and protein appear in 
the smaller ribosomes (30S and 50S), and 
these particles are the precursors of the 
larger (70S and 85S) ribosomes. Further- 
more, there is evidently a rapid circulation 
of material among the various classes so 
that the 70S and 85S particles break down 
to form 30S and 50S particles. Thus the 
different groups of particles seem to rep- 
resent different phases of the growth cycle 
of the ribosomes. This interpretation was 
suggested in CIW Year Book 56 (1956- 
1957), but at that time there were no 
experimental facts to support it. 

The mechanism of protein synthesis can 
also be investigated by a study of the 
errors that can occur. When certain amino 
acid analogs (compounds very similar to 
the usual amino acids of the cell) are sup- 
plied to a growing culture of cells, they 
may be incorporated in place of the cor- 
responding natural amino acid. The con- 
ditions that give rise to these mistakes and 
the effect of them on the resulting proteins 
provide useful information both as to 
which elements of the cell constitute the 
mechanism to select the proper amino acid 
and as to which elements of a polypeptide 
strand are essential for enzymic activity. 

This year has brought a better under- 
standing of some of these processes. When 
cells are supplied with an amino acid 

and its analog, both are incorporated, but 
the cell usually discriminates against the 
analog. Hence, the ratio of analog to 
amino acid is very much less in the newly 
formed protein than in the original sup- 
ply. Measurements carried out with yeast 
have shown that the internal pool of 
amino acids has the same ratio of analog 
to amino acid that occurs in the protein. 
There seems to be no further selection in 
those steps that transfer the amino acids 
of the pool into the final proteins. Con- 
sequently, selection occurs in those proc- 
esses by which the amino acids enter the 
pool, and less specificity is required in the 
template which arranges the amino acids 
into proper order. Alternatively the amino 
acids found in the internal pool may al- 
ready have been selected by the template 
but not yet linked into polypeptide strands. 

In addition to the studies of microorgan- 
isms carried out in this laboratory, we have 
continued to cooperate with Drs. L. B. 
and J. B. Flexner (Department of Anat- 
omy, University of Pennsylvania) in their 
studies of the development of the brain. 

The details of the experimental work 
and a more complete discussion of its 
interpretation will be found below. 


Composition of Cell 

A number of structures that seem to be 
related to the many different functions of 
the growing cell have been recognized in 
microorganisms. In E. coli a cell wall, a 
cell membrane, soluble protein, soluble 
RNA, ribosomes, and DNA can be dis- 
tinguished. The cell wall provides struc- 
tural rigidity. It can be weakened, or in 
some species dissolved away, leaving a 
fragile "spheroplast" or "protoplast" hav- 
ing unimpaired synthetic capacities. The 
cell membrane which lies next to the wall 
contains a large part of the lipid of the 
cell. Its function is less clearly known; it 
is probably important in the processes by 
which small molecules are transported 
into the cell, it may serve to hold some of 


the soluble protein in a spatial association 
which permits organized enzymatic op- 
eration, and it may serve as a site of at- 
tachment for ribosomes. Together the 
wall and membrane make up roughly 25 
per cent of the dry weight of the cell, 
including one-sixth of the protein and 
most of the lipid. 

DNA (3 per cent of the cell mass) is 
the genetic material. It provides the in- 
formation for its own duplication and 
may participate in the synthesis of RNA 
and protein. 

The soluble proteins that carry out most 
of the enzymatic functions of the cell 
account for roughly 40 per cent of the 
cell and two-thirds of the protein. Roughly 
5 per cent of the cellular material is soluble 
RNA, which is considered by some in- 
vestigators to have a role as an amino acid 
carrier in protein synthesis. The quantity 
of soluble protein per cell is equivalent 
to roughly 6 million units of molecular 
weight 10,000. At the usual growth rate 
of 2 per cent increase per 100 seconds, 
1200 of these units must be synthesized 
per second per cell. 

The ribosome content varies consider- 
ably, depending on the growth condition, 
being highest during rapid growth. Ribo- 
somes account for 15 to 30 per cent of the 
dry cell weight, for 50 to 85 per cent of 
the RNA, and for 15 to 30 per cent of the 
protein. As will be shown below, the 
ribosomes provide the principal machinery 
for protein synthesis. The ribosomes occur 
in a number of different sizes; in the 
growing cell, ribosomes of sedimentation 
constants 15 to 20S, 30S, 50S, 70S, and 85S 
are usually observed. The distribution 
among the different classes indicates that 
there are roughly equal numbers of each, 
giving about 2000 ribosomes of each size 
per cell. Thus, one new ribosome of each 
size must be synthesized every 2 l / 2 seconds. 

As the number of different enzymes 
required to carry out the different reac- 
tions of the cell seems to be of the order of 
1000, there appear to be enough ribosomes 

to provide one for the synthesis of each 
different enzyme. Furthermore, if the 70S 
ribosomes synthesize the soluble protein, 
and if all are equally active, then each 
ribosome would synthesize about 1 unit 
of molecular weight 10,000 per 2 seconds. 

Analytical Methods 

Any investigation of ribosomes depends 
on methods for separating them from 
other cellular material and separating the 
various classes of ribosomes. Thus, to 
show that one class of ribosome is the 
precursor of another the two classes must 
be separated and purified. Similar analyti- 
cal methods are needed to show whether 
enzymes are associated with ribosomes, 
whether analogs enter one or another pro- 
tein. As these methods are not yet stand- 
ardized, it is necessary to give a brief 
description of them and their limitations. 
In general, separation techniques are based 
on differences in sedimentation rate, in 
electrophoretic mobility, in affinity to ab- 
sorbing materials, or in chemical proper- 
ties such as solubility and resistance to 

Sedimentation. The analytical centri- 
fuge is the most convenient and precise 
tool for showing which groups of particles 
are present in cell juices or for analyzing 
fractions obtained from other procedures. 
Unfortunately, the components which it 
resolves are not readily available for meas- 
urement of radioactivity or composition. 

Differential centrifugation provides a 
useful separation of components of quite 
different sedimentation constants. Cell 
walls and membranes can be removed by 
5 minutes' centrifugation at 40,000 rpm 
(40K 5') in the angle head of the Spinco 
model L. Slightly longer centrifugation 
(40K 30') yields a pellet rich in the large 
ribosomes (70 to 85S) whereas prolonged 
centrifugation (40K 120') will remove al- 
most all the ribosomes. These procedures 
are most useful in removing cell walls 
or in separating ribosomes from soluble 
RNA and protein. They do not give an 



adequate separation of the different classes 
of ribosomes. 

It is possible, however, to separate the 
different classes by centrifuging a layer of 
ribosomes through a sucrose density gra- 
dient, and a degree of separation was 
achieved last year. This method has been 
enormously improved as a result of an 
examination of the conditions of stability 
of a liquid column in a gravitational (or 
centrifugal) field. 

Common use has been made of density 
gradients for the stabilization against mix- 

ure 10A, the system will be initially stable. 
As soon as the sample is moved down- 
ward, however, the possibility of insta- 
bility arises at the lower edge of the sam- 
ple band. If the sample makes a large 
contribution to the liquid density, or has 
not been spread sufficiently by diffusion, 
the inverted density gradient due to the 
sample (at point X) will exceed the sta- 
bilizing gradient and the analysis will be 
spoiled by streaming down of the sample. 
If the sample is introduced on top of 
the liquid column, not in a uniform layer, 

Increasing liquid density >• 

A unstable at X 

B stable 

Fig. 10. Conditions of stability in a density gradient liquid column. The solid line represents 
the stabilizing gradient and the dotted line the contribution due to a sample in process of analysis. 

ing of liquid columns for zone (or differ- 
ential) analysis by means of centrifugation 
or electrophoresis. This method has been 
limited to very small quantities of ma- 
terial, however, since the sample itself 
may introduce a region of instability. So 
long as the density increases in the direc- 
tion of the gravitational (or centrifugal) 
field the gradient will exercise a stabiliz- 
ing force against mixing due to mechani- 
cal disturbances or temperature gradients. 
When the density gradient is reversed a 
condition of instability exists and usually 
the upper liquid streams rapidly through 
the less dense underlying layers. 

When a sample is loaded onto the top 
of a liquid column having an artificially 
created density gradient, as shown in fig- 

but with a gradient in concentration op- 
posite to that of the stabilizing gradient 
(fig. 105), samples of indefinitely large 
size can be analyzed. The condition must 
be met, however, that the inverted gradi- 
ent in density due to the sample be sig- 
nificantly less than the stabilizing density 

The amount of sample that can be 
handled rises as the square of the width 
of the zone in which it is initially loaded 
since it is the gradient in density and not 
the maximum density of the sample that 
determines the stability. As a result the 
maximum attainable resolution falls with 
the square root of the quantity of material 
in the sample to be analyzed. 

This principle has been applied success- 


fully for analysis both by electrophoresis 
and by sedimentation in the ultracentri- 
fuge. Since analysis by sedimentation has 
been of great importance to our experi- 
ments on the synthesis of and by the 
ribosomes of E. coli it will be described 
in detail. 

The linear stabilizing density gradient 
was produced by means of sucrose (20 to 
5 per cent) dissolved in the appropriate 
buffer solutions. The required concentra- 
tion change was achieved with the linear 
gradient mixing device of Bock and Ling 
(fig. 11). Such gradients are stable for 

platinum stirrer 



Fig. 11. Mixing chamber for the production 
of linear stabilizing gradients. The left chamber 
is filled with 2.4 ml 5 per cent sucrose, and the 
right with 2.2 ml 20 per cent sucrose. After the 
chambers are filled and the mixing motor is 
started the center valve is opened and the exit 
tubing is turned down to touch the side of the 
centrifuge tube. The sucrose solution then runs 
down the wall of the tube. If the 4.6 ml are 
delivered in 10 to 15 minutes the succeeding 
lighter solution floats on the underlying liquid 
with little mixing. 

many hours, and the tubes can be handled 
with only moderate care. The same device 
was then used to introduce the inverted 
gradient of the sample to be analyzed. 
For this purpose the left-hand chamber 
was loaded with 0.2 ml (for example) 

of buffer containing 5 mg of ribosomes 
per milliliter, and the right-hand chamber 
with 4 per cent sucrose in the same buffer. 
The sudden step in sucrose concentration 
from 5 per cent to 4 per cent allows the 
sample gradient to be started without un- 
due mixing. The use of a separate and 
still smaller mixing device for sample load- 
ing would probably be justified. 

This whole process was carried out in 
the cold room with solutions at 1° to 4° C; 
the tube was loaded into the precooled 
swinging-bucket rotor (SW-39) and cen- 
trifuged for the appropriate time in the 
refrigerated model L preparative ultra- 
centrifuge. As quickly as possible at the 
end of the run the tube was gently lifted 
out of the rotor and mounted in a device 
which perforated the bottom of the cen- 
trifuge tube with a hypodermic needle 
ground to a short double-sided point, 
which was located about 1 mm above the 
bottom of the tube. After removing a 
piano wire which kept the hypodermic 
tubing clear and free of air bubbles the 
contents of the tube were run out in 25 
equal cuts by drop counting. 

In our hands a sample width such as 
described above where the sample was 
initially graded from zero to full concen- 
tration over 0.4 ml has given the best 
compromise between high resolution and 
a useful quantity for analysis. The resolu- 
tion could probably be improved to the 
limit set by diffusion with great care in 
the process of sample loading. For the 
preparation of large quantities of material 
very broad sample gradients are useful. 
For example, in the preparation of pure 
30S particles from a mixture of 30S and 
50S ribosomes, 2.4 ml of a very concen- 
trated ribosome suspension can be placed 
in the left chamber and 2.4 ml of 20 per 
cent sucrose in the right chamber. In this 
way 10 to 20 mg of pure 30S particles can 
be obtained in a single run with a yield 
of 30 to 50 per cent (see fig. 12, pi. 2). 

It is worth noting that while the centrif- 
ugal force doubles from the top to the 


bottom of the tube in the swinging-bucket 
rotor the viscosity of 20 per cent sucrose 
is about twice that of water. When cor- 
rection is made for the density of the 
sucrose solution it is found that the sedi- 
mentation velocity of ribosomes is very 
nearly constant throughout the length of 
the centrifuge tube under these conditions. 
A number of examples of the application 
of this technique appear in other sections 
of the report. 

Chromatography. Chromatography on 
modified cellulose (DEAE) columns con- 
tinues to be a most useful analytical 
method. The material to be analyzed is 
adsorbed by the column and then eluted 
with a salt gradient. Examples of its reso- 
lution of different proteins are shown in 
figure 30. Unfortunately the resolution of 
the nucleic acids and nucleoproteins is not 
quite so sharp. Four regions on the elu- 
tion diagram which contain nucleic acid 
may be distinguished. These have been 
designated A, B, C, and D in the order of 
their elution (fig. 13). 

Purified ribosomes give rise to peak A; 
its position is unchanged, whether 30, 50, 
or 70 to 85S particles are adsorbed. The 
eluted material has only half the protein 
of the ribosomes, whether purified 30S or 
50S particles or a mixture of all particles 
is analyzed. The sedimentation constant 
of 30 or 50S particles is not appreciably 
changed as the increased density of the 
resulting nucleic acid-rich particle com- 
pensates for the decrease in mass. The 
70 and 85S particles emerge as a mixture 
of the protein-deficient 30 and 50S par- 

Soluble RNA gives rise to the B peak 
of the elution diagram. It is not removed 
from cell juice by centrifugation, and there 
is little if any associated protein. 

A peak designated C may appear in cell 
juices containing DNA. Some of the ma- 
terial eluting in the C region remains pre- 
cipitable after hydrolysis with KOH. In 
most preparations the DNA is degraded 
by the pressure cell and by the DNAase 

of the cell juice, and a well defined C 
peak does not appear. 

Another peak designated D contains 
most of the radioactivity incorporated into 
nucleic acid during a short exposure of 
cells to P 32 C»4 = . It contains little ultraviolet- 
absorbing material, however, except when 
the A peak is collected and rechromato- 
graphed, or when a ribosome preparation 
has aged for several days, or when the 
cells are incubated in the presence of 
chloramphenicol for 2 hours. 

lr. 10 

30 35 40 45 50 55 

Fraction number 

Fig. 13. DEAE-cellulose chromatography of 
nucleoproteins and nucleic acids of E. coli. Re- 
gion A is nucleoprotein derived from ribosomes; 
B is soluble RNA; C is DNA; and D is derived 
from precursor nucleoprotein. The solid line 
indicates the optical density at 260 u, and the 
dotted line shows P 32 which the bacteria had 
incorporated during a 2-minute labeling period. 
The NaCl concentration in fraction 40 was 
0.45 M. 

The behavior of particles on the column 
is strongly dependent on the magnesium 
content of the eluting fluid. When mag- 
nesium is omitted the A and D peaks are 
not eluted by the 0-1 M gradient of NaCl. 
The omission of magnesium does not, 
however, influence the elution of the B 
peak. When the magnesium concentra- 
tion is increased tenfold to 0.05 M the 
A and D peaks are eluted at a lower NaCl 
concentration. Again there is no change 
in the elution of the B material. Once 

Plate 2 

Department of Terrestrial Magnetism 

Fig. 12. Analytical ultracentrifuge pattern of 
purified 30S ribosomes. E. coli ribosomes were 
prepared by repeated centrifugation in buffer 
containing 10~ 4 M MgAc yielding a mixture con- 
taining equal numbers of 30S and 50S derived 
ribosomes. The 30S particles were then purified 
by means of the swinging-bucket-sucrose-gradient 
technique described in the text. 

Fig. 14. Ribosomal pattern of cells of E. coli 
exponentially growing in broth. Cells washed 
and broken in TS at 10~ 2 M Mg. Schlieren plate 
taken 5 minutes after the analytical ultracen- 
trifuge had reached 56,740 rpm. The peaks are 
labeled with nominal sedimentation constants. 
In comparison, cells grown in a salts-glucose 
medium show relatively less of the 70 and "85" 
S peaks. 

Plate 3 

Department of Terrestrial Magnetism 

Fig. 19. Sedimentation diagrams of bacterial juices as a culture of E. coli is brought from the 
resting to the growing state by the addition of glucose. A to F are the diagrams at 0, 1/4, 3, 7, 
11, and 15 minutes after adding glucose. 

Fig. 20. Sedimentation diagram of E. coli juice after the cells had been magnesium-starved overnight. 



adsorbed to the column in the absence of 
magnesium, the A material cannot be 
eluted even though magnesium is added 

Since most proteins have less affinity 
for the column than nucleic acid has, it 
would seem that the particles are held to 
the column by their nucleic acid moiety 
(i.e., the nucleic acid is exposed and not 
completely covered, as in some viruses). 
The binding of nucleic acid to the column, 
however, does not vary with the magne- 
sium concentration. Hence, it seems quite 
possible that the effect of magnesium on 
the binding of the particles comes about 
through an effect on their shape. Accord- 
ing to this picture, in low magnesium, the 
particles are able to unroll into long 
strands which are capable of making many 
more bonds to the column. 

Electrophoresis. A few studies of elec- 
trophoresis of the ribosomes were carried 
out since electrophoresis offered a possi- 
bility of separating the macromolecular 
precursors of the ribosomes which had 
been identified by DEAE column analysis. 
In fact, no separation was achieved, sug- 
gesting that the precursors were very 
similar to the ribosomes themselves but 
differed in the lability during DEAE col- 
umn chromatography. 

Two methods were used for these stud- 
ies. The first (which supplied somewhat 
higher resolution but was less convenient) 
made use of a density-gradient-stabilized 
liquid column similar to that described 
in the previous section. A 30-cm-high 2 x / 2 - 
cm-diameter liquid column with a stabiliz- 
ing gradient of glycerol (50 to 5 per cent) 
was prepared in a glass tube with a nylon 
filter cemented over the lower end. The 
upper end was closed with a rubber stop- 
per through which passed a U tube sealed 
off with a Visking membrane to supply 
the cathode connection. The whole tube 
was immersed in a large cylinder contain- 
ing the platinum anode, and the external 
liquid level was adjusted to match the 
internal level. After electrophoresis at 6 
ma for 18 hours in the cold room the tube 

was lifted out of the external cylinder and 
the contents were dripped out into a frac- 
tion collector. 

For most of the runs the same mechani- 
cal arrangement was used, but the glycerol 
stabilizing gradient was replaced with a 
packing of Geon 427 polyvinyl resin. The 
Geon 427 is a very successful packing since 
it consists of small spheres of nearly uni- 
form size and has no affinity for the ribo- 
somes and very little endosmotic flow. 
With the resin packing it was necessary 
to elute at slow flow rates over a period 
of 1 or 2 hours to avoid excessive spread- 
ing of the bands. All the electrophoretic 
runs were made in the standard buffer 
(TSM) at pH 7.6. 

Although the ribosome peak was fairly 
broad (2 cm wide after 17 cm migration) 
and not fully resolved from the soluble 
proteins of the cell, it was possible to de- 
tect small differences in electrophoretic 
mobility by running various labeled prep- 
arations mixed with an excess of unlabeled 
ribosomes, and measuring the specific ra- 
dioactivity across the ribosome peak. 

When P 32 pulse-labeled juice (showing 
almost all the radioactivity in the D region 
on DEAE chromatography) was run, no 
difference in electrophoretic mobility was 
observed between the precursors and the 
mature ribosomes. The specific radioac- 
tivity across the ribosome peak was con- 
stant within 15 per cent. 

In another test the same P 32 pulse- 
labeled juice was first analyzed on the 
DEAE column and the D region of the 
column elution pattern was centrifuged 
for several hours at 105,000^. When a 
mixture of this pellet (column-degraded 
precursor) and unlabeled ribosomes (not 
subjected to column degradation) was an- 
alyzed by electrophoresis the radioactivity 
showed a 7 per cent greater mobility than 
the ribosomes. The specific radioactivity 
across the ribosome peak varied by a fac- 
tor of 5. 

For the interpretation of these results 
in terms of protein content of the pre- 
cursor a mixture of unlabeled ribosomes 


and labeled nucleoprotein eluted from die 
column (CNP) was analyzed. This col- 
umn nucleoprotein has half the quantity 
of protein per unit nucleic acid as the 
original ribosomes. In this case the CNP 
moved somewhat faster than the ribo- 
somes (~3 per cent) and the specific 
radioactivity across the ribosome peak 
varied by a factor of 2. 

As a further check the juice of cells 
labeled with P 32 in the presence of 50 
ug/ml of chloramphenicol was analyzed. 
Under these conditions labeled RNA is 
synthesized but no nucleoprotein is 
formed. Upon electrophoresis the bulk 
of the radioactivity moved out in two 
peaks ahead of and well resolved from the 
ribosomes, and a new RNA peak was 

It is clear that the precursors before 
column degradation migrate identically to 
the ribosomes and after column degrada- 
tion migrate faster than the protein-poor 
column-degraded nucleoprotein, but slower 
than the normal soluble RNA of the cell 
or the RNA produced in the presence of 

A definitive interpretation of these re- 
sults in terms of the protein content of 
the precursors is not possible since on the 
one hand the effect of the protein content 
on the mobility appears to be small at this 
pVL and on the other hand properties other 
than the protein content may influence the 

It appears likely, however, that the pre- 
cursors are initially similar to the ribo- 
somes in protein content but on column 
degradation lose a greater fraction of their 

Chemical fractionation. Chemical frac- 
tionation is entirely adequate for separat- 
ing different classes of compounds and is 
needed to measure the composition of vari- 
ous fractions obtained by other separation 
procedures. It has been less useful in sepa- 
rating the macromolecules of cell juices. 
Precipitation with 2-chloroethanol may be 
of some value in making a crude separa- 
tion between ribosomal particles and the 

soluble materials of the cell juice. The 
addition of 2-chloroethanol to cell juice so 
as to obtain a final concentration of 5 per 
cent results in the precipitation of most 
of the ribosomes and very little of the 
soluble RNA or protein. As the final con- 
centration of the reagent is increased from 
6 to 13 per cent more of the soluble pro- 
tein is precipitated. At concentrations of 
more than 15 per cent the cell juice fails 
to give any precipitate. Such a crude frac- 
tionation between ribosome and "soluble" 
makes possible rapid analysis of the par- 
ticle content of a cell juice. 

Sedimentation Constants of Ribosomes 

In the previous annual report (Year 
Book 57) the four major sizes of ribo- 
somes were described (sedimentation con- 
stants approximately 20, 40, 60, and 80S) 
and also the magnesium requirement for 
their stability. During the past year nearly 
400 runs have been made with the ana- 
lytical ultracentrifuge both to monitor 
preparations for different types of experi- 
ments and for further study of the nature 
and stability of the ribosomes. 

The previously quoted apparent sedi- 
mentation constants were observed in total 
cell juices. Correction for the viscosity of 
the cell juices and studies of purified ribo- 
some preparations showed that there are 
ribosomes of sedimentation constants of 
about 20, 30, 50, 70S and a larger com- 
ponent which varied in sedimentation con- 
stant from just over 70 to about 100S de- 
pending on the magnesium concentration 
and the growth condition of the bacteria. 
These sedimentation constants have been 
rounded off for convenience in discussion. 
Exhaustive studies of sedimentation as a 
function of concentration have not been 
made. The concentration dependence is 
small, however, and these figures are ac- 
curate to ±5 per cent and consistent with 
published values for purified ribosomes. 

Figure 14, plate 2, shows the Schlieren 
diagram of a total juice from exponentially 
growing cells. In this experiment the cells 



were washed three times in TSM (10~ 2 M 
Mg). This magnesium concentration is 
high enough to drive the equilibrium 
among the various forms of ribosomes over 

have very different specific radioactivities 
from the larger forms. Thus, these small 
forms are not simply dissociation frag- 
ments from the larger ribosomes and are 


4.33 x 10" 


4.83xl0" 3 92.0 

3.33 x I0" 3 


3.83 x 10 





1.33 x 10 







Fig. 15. Ribosomal pattern as a function of magnesium concentration. Tracings of 10X enlarged 
prints of Schlieren plates taken 5 minutes after the analytical ultracentrifuge had reached 50,740 rpm. 
The numbers at the left are the calculated magnesium concentrations in moles per liter. The num- 
bers over the peaks are the measured sedimentation constants, corrected only for the viscosity change 
of water to 20° C. Ribosome preparation described in text. 

toward the larger (70 and 100S) forms, 
but sizable quantities of small ribosomes 
(20, 30, 50S) still remain. A number of 
tracer experiments have shown that these 
smaller, naturally occurring particles may 

identified for purposes of discussion as the 
small "native" ribosomes. 

Figures 15, 16, and 17 show the results 
of a study of the effects of magnesium 
concentration on the quantity and sedi- 



mentation constants of the various forms 
of ribosomes. Exponentially growing 
(broth) cells were washed three times in 
TS containing 2xl0 -2 M MgAc, broken 
in the pressure cell, and centrifuged to 

e 60 

Fig. 16. Variation of sedimentation constant 
of the various forms of ribosomes with mag- 
nesium concentration. Data from figure 15. 











II \ _ 

II \ 

II \ 




70S " — *- 





i~V~*— r-r 


, 30S L 

10"' 10" 2 

Mg Ac cone, mols/liter 

Fig. 17. Variation in the quantity of the vari- 
ous forms of ribosomes as a function of mag- 
nesium concentration. Data obtained from the 
Schlieren diagrams of figure 15 simply by cutting 
out and weighing the peaks from enlarged prints. 

remove cell bodies and wall and mem- 
brane fragments. Then the larger ribo- 
somes were harvested by centrifuging 45 
minutes at 105,000g-. These were then 
resuspended in a small volume of the 

same buffer (TS 2X10" 2 M Mg). Small 
samples of this concentrated ribosome sus- 
pension were diluted a factor of 14 (to 
3 mg RNA per ml) into TS to give 
the magnesium concentrations indicated. 
These calculated concentrations are not 
corrected for the magnesium due to dis- 
sociation of the Mg-ribosome complex. 

The major observations of the effect of 
magnesium on the ribosomes may be sum- 
marized as follows: (1) Decreasing the 
magnesium concentration from 6 X 10~ 3 
to 2 X 10" 3 M converts the 80 to 100S form 
into the 70S form without the appear- 
ance of any significant quantity of 50 and 
30S forms. (2) At concentrations below 
2 X 10" 3 M Mg the 70S ribosomes dissoci- 
ate into 50 and 30S forms. (3) Removal 
of all magnesium (and other polyvalent 
cations) with EDT dissociates the 50 
and 30S ribosomes in turn, with the 
consequent digestion of their RNA by 
the RNAase contained in the ribosomes. 
(4) The sedimentation constant of the 
70S form is slightly dependent on the 
magnesium concentration below 3 X 10 -3 
M Mg. (5) The sedimentation constant 
of the 80 to 100S form is strongly depend- 
ent on the magnesium concentration be- 
tween 2 and 6xl0" 3 M Mg. (6) The 
peak of variable sedimentation constant 
is noticeably broader at the intermediate 
magnesium concentrations. (7) The 70S 
and 80 to 100S forms exist in equilibrium 
with each other since reduction of the 
magnesium concentration from 2 X 10 -2 M 
to 1.8 X10~ 3 M and restoration to 4.33 X 
10" 3 M yields the same picture (after over- 
night storage at 4° C) as was obtained 
by direct dilution of the magnesium con- 
centration to 4.33 X10" 3 M. (8) Dilution 
of the ribosome concentration (to 1 mg 
RNA/ml) does not change the relative 
quantity of the 70 and 90S peaks at 4.33 X 
10~ 3 M Mg either immediately or after 
storage overnight at 4° C. 

It appears that the 70S particle is a com- 
bination of 30S and 50S forms in equal 
numbers. If the 30, 50, and 70S particles 
were each approximately spherical and of 


the same hydration and density, one 30 
and one 50 might form a 70S particle. 
On the other hand, if the 70S form were 
a relatively more extended object it is 
possible that it could consist of two 30S 
and two 50S particles. The choice between 
these possibilities cannot yet be made with 

Again, if the 70S were spherical and 
two 70S could dimerize to form a sphere 
of similar hydration and density the re- 
sultant particle would have a sedimenta- 
tion constant of about 110S. Standing 
against this, however, are items 5 and 8 
in the list above. The smooth change of 
sedimentation constant with magnesium 
concentration suggests an object of mag- 
nesium-dependent shape or hydration. The 
shape change required is quite striking 
since the transformation of a sphere to 
an ellipsoid of axial ratio 6 of identical 
mass and hydration would be necessary 
to reduce the sedimentation constant by 
about 30 per cent. 

An alternative interpretation is that the 
intermediate values of sedimentation con- 
stant measure the average sedimentation 
rate of objects that are rapidly changing 
from a 70S to a 100S form and indicate 
the fraction of time the particle spends in 
each of the states. The presence of a 70S 
peak as the magnesium concentration is 
lowered through the critical range requires 
an additional hypothesis that the 70S par- 
ticle exists as well in a state incapable of 
making the transformation. If a rapid 
transformation were due to a dimerization 
reaction one would expect that the equi- 
librium would be shifted in favor of the 
70S form by reduction of the ribosome 
concentration. Since this is not observed 
it appears that variations of shape or hy- 
dration have a major effect on the sedi- 
mentation constants of the ribosomes and 
may even totally account for the transition 
from the 100S to the 70S form. Attempts 
to settle this question by means of light- 
scattering measurements of the molecular 
weights have not yet been successful, ow- 
ing to the presence of small quantities of 

wall or membrane fragments of enormous 
molecular weight but of sedimentation 
constants in the 70 to 100S range. 

On figure 17 is shown a theoretical 
curve (dotted) representing the quantity 
of the 80 to 100S form expected for a 
transformation which depends upon the 
fifth power of the magnesium concentra- 
tion. The implication of the good fit with 
the experimental data is that roughly 5 
Mg atoms are required to trigger the 
transformation. Other published data in- 
dicate that there may be 1000 Mg atoms 
total absorbed per ribosome. If the shape 
change were a result of the formation of 
magnesium bonds which caused cross link- 
ing and contraction of an initially fairly 
open structure it is reasonable to suppose 
that a few magnesium atoms in a critical 
region could initiate such a transforma- 
tion by drawing the strands together and 
thus allowing other cross-linking bonds 
to be formed sequentially down the length 
of the strands. 

The relative quantity of 70 and 80 to 
100S particles depends on the growth con- 
dition of the cells. The quantity of nas- 
cent protein associated with the 70 and 
80 to 100S particles is observed to be dif- 
ferent. These results imply that the trans- 
formation between these two forms is in- 
timately associated with the growth of 
the cell and the synthesis of protein. A 
shape change of the ribosomes might be 
useful in freeing newly synthesized pro- 
tein from the intimate association with the 
ribosomes that must exist during synthesis. 


The ribosome content is characteristic 
of the physiological state of the cell, both 
in its magnitude and in the relative quan- 
tities of the various component classes. 
The absolute quantity of particles in a 
cell varies according to the medium in 
which they are growing, being greater 
when the cells are growing more rapidly. 
Furthermore, radical changes in the rela- 
tive proportions of the various sizes are 


discernible as the cell passes from the 
growing to the resting state, and vice 
versa. The latter variations are quite differ- 
ent from those brought about by changes 
in the magnesium-ion concentration de- 
scribed above, since all washing and break- 
ing are carried out in the standard TS2M 
buffer. An attempt has been made to 
interpret these relationships considering 
the 70 and 85S components to be the sites 
of protein synthesis so that their absolute 
quantity determines the rate of protein 
synthesis and consequently the over-all 
rate of growth (85S is used to indicate 
the particle of variable sedimentation con- 
stant ranging from 70 to 100S). 

There do not seem to be any major 
changes in either the quantity of particles 
or the distribution among the various 
classes during the growth cycle of an in- 
dividual cell. The ultracentrifuge patterns 
of samples taken through the division 
cycle of a synchronous culture of cells of 
Alcaligenes fecalis are essentially identical 
(fig. 18). The quantity of particles rela- 
tive to the soluble protein peak is constant. 
The distribution between the various par- 
ticle sizes is identical to that in a random 
culture. This would seem to indicate that 
the synthesis of particles is virtually con- 
tinuous throughout the division cycle and 
not confined to a minor part. Studies of 
ribosomes made in an unsynchronized cul- 
ture are applicable to the growth of single 
cells. This experiment was carried out in 
cooperation with Dr. K. G. Lark, who 
supplied the samples of the synchronized 

Exponentially growing cells have a par- 
ticle pattern dominated by the larger ribo- 
somes of 70 and 85S. The smaller particles 
of 50, 30, and 20S are also present in 
appreciable quantities. On the other hand, 
in resting cells which have run out of the 
energy source, glucose, the particles are 
almost entirely of a single size (100S). 

More detailed examination of samples 
taken from a culture undergoing the 
transition from the resting to the growing 
state has shown that these changes in par- 

ticle pattern proceed with great rapidity. 
Centrifugal analysis of samples taken dur- 
ing the first 15 minutes after addition of 
glucose to a culture which had exhausted 
its glucose supply about 16 hours pre- 
viously showed that appreciable conversion 
of the 100S to the 70, 50, and 30S particles 
had occurred during the first V/ 2 minutes 
(fig. 19, pi. 3). The conversion progressed 
to more than 50 per cent by 7 minutes. 
At 11 minutes all the 100S component 

Fig. 18. Sedimentation diagrams of cell juices 
of bacteria (Alcaligenes fecalis) undergoing divi- 
sion in synchrony. The seven samples were taken 
serially during the course of a cycle of division. 

had disappeared and the predominant par- 
ticle was one with a sedimentation con- 
stant of 88S (table 14). At this time the 
pattern resembled closely that of a typical 
exponential culture. Study of the S 35 Ov 
uptake showed that this period between 
7 and 11 minutes also corresponded to a 
period of increasing rate of protein syn- 
thesis which reached a maximum at 12 

A similar study suggested that changes 
in particle pattern are equally rapid dur- 
ing the reverse transition into the resting 
state. Although there is more uncertainty 


in assessing the time at which the last 
traces of glucose are swept up by the cells, 
the centrifugal analysis of two samples 
taken 5 minutes apart just as the turbidity 
of the culture reached a plateau gave two 
very different patterns. Whereas the first 
was typical of an exponential culture, the 
second had none of the 85S particles and 
showed a predominant 100S peak. 

In view of the observed interconversion 
of the 70 and 85S components it is perhaps 
questionable whether their relative pro- 
portion has any physiological significance. 
There is always the possibility that transi- 
tions could occur as artifacts of the wash- 
ing or breaking procedure. The best evi- 

TABLE 14. Changes in Ribosomes upon Re- 
sumption of Growth of E. coli 

Minutes after 

Addition of 


S 20 values 



104, 75, 50 
98, 70, 54 
100, 74, 57 
86, 70, 56 
85, 69, 52 

dence that these objects are distinct in the 
cell, or are at least derived from different 
objects, comes from the different specific 
activities of the two peaks after a short 
S 35 04 = pulse (fig. 33). It is not unreason- 
able, therefore, to assume that each of the 
three 70, 85, and 100S components are 
distinct objects, the relative proportion of 
which has a close correlation with the cell 

All these observations are consistent 
with the view of the 70 and 85S compo- 
nents as the active machinery in protein 
synthesis and the more rapidly sediment- 
ing 100S particles as an inert form. Transi- 
tions between these three particle types 
occur freely in the cell as well as in the 
test tube. Since the test-tube transforma- 
tions are effected by means of Mg ++ ion, 
it is possible that in vivo changes might 

be made by means of a magnesium-con- 
centrating mechanism. 

Detailed accounts have already been 
given of the importance of magnesium 
in the preservation of ribosomal structures 
in vitro. Complementary studies have been 
made on the effect of the lack of magne- 
sium upon cell growth and on the particle 
content of cells. When growing cells, 
washed free of magnesium ions, are in- 
oculated into a magnesium-free medium, 
growth proceeds for a while on the traces 
of magnesium and then ceases. Further 
periods in this deficient medium cause a 
gradual reduction in the number of par- 
ticles present in the cells. The particle 
content of cells left overnight may be re- 
duced to about 5 per cent of normal (fig. 
20, pi. 3). 

Very little loss into the medium of 260- 
m|j absorbing material occurs under these 
conditions, and the ribosomal RNA re- 
mains TCA precipitable within the cell. 
This degenerate material is eluted in the 
D region on the DEAE column. Sedi- 
mentation analysis in the swinging-bucket 
rotor proved that it has an S number of 
less than 20. 

An experiment in which the ribosomal 
RNA was prelabeled with P 32 Gv, before 
overnight magnesium starvation, showed 
that during subsequent growth in P 31 Gv 
(after addition of magnesium) the degen- 
erate ribosomal RNA was not reassembled 
into particles but appeared to be slowly 
broken down into TCA-soluble pool ma- 
terial. The over-all effect of magnesium 
starvation is then the irreversible breakup 
of the ribosomes into the protein and 
RNA components. Subsequent recovery 
proceeds by de novo synthesis of at least 
the ribosomal RNA if not the protein also. 

Further attention was paid to recovery 
from magnesium starvation with a view 
to studying the relationship between par- 
ticle content and the rate of growth and 
protein synthesis. Growth resumes soon 
after the addition of magnesium but at a 
very slow rate. The growth rate increases 
continuously, although complete recovery 


from overnight magnesium starvation rate 
may require many hours (fig. 21). 

In one such experiment the rate of pro- 
tein synthesis was measured by means of 
S 35 4 = incorporation. In addition to fol- 
lowing the optical densities of the culture 
at 650 mu and 260 mu, large samples of 
the culture were taken at intervals, washed 
and broken, and analyzed for particle con- 
tent by centrifugation. Since the 260-m[j 
and 650-m|j optical densities proved to be 
proportional throughout the experiment, 
the percentage of the 260-m|j absorbing 


Mg ++ added 

J I I L 

J I L 

3 4 


Fig. 21. The effect of magnesium restoration 
on the growth of magnesium-starved bacteria. 

material which sedimented (40K 90') was 
taken as an adequate measure of the rela- 
tive particle content. The rate of protein 
synthesis per unit mass of cells was di- 
rectly proportional to the ribosome content 
(fig. 22). 

The increase in the total number of 
particles seems to be autocatalytic under 
these nonsteady-state conditions. The total 
number of particles increases exponentially 
by a factor of more than 80 (fig. 23) . The 
mean generation time of the ribosomes (50 
minutes) is approximately that of the cells 
growing under steady-state conditions in 
the same medium. As the mean genera- 
tion time of the cells is initially much 

longer, the ribosome content per cell is 
eventually restored to its normal value. 
This suggests that the ribosome synthesiz- 
ing system may be a reasonably auton- 
omous one. 

Another experiment has shown a rela- 
tionship between DNA and ribosome syn- 
thesis. The thymine-requiring mutant of 
E. coli, 15T~, cannot synthesize DNA but 
does continue to synthesize RNA when 
thymine is lacking. The incorporation of 
P 32 into soluble RNA and into ribosomes 
was compared in 15T" with thymine pres- 
ent or absent. When the quantity of 

1 2 3 4 5 6 7 0.8 

Rote of protein synthesis (Arbitrary units) 

Fig. 22. The relationship between rate of 
protein synthesis as measured by S 35 incorpora- 
tion and the ribosome content of E. coli recover- 
ing from magnesium starvation. 

DNA stopped increasing because of the 
thymine deficiency, the rate of ribosome 
synthesis stopped increasing and thereafter 
remained constant even though the quan- 
tity of ribosomes had doubled. In con- 
trast, the rate of soluble RNA synthesis 
continued to rise. The analytical centri- 
fuge and chromatography on DEAE in- 
dicated that the ribosomes formed during 
this period were normal; the only differ- 
ence observed was that there were more 
20, 30, and 50S particles in the thymine- 
deficient cells than in the control. 
In view of the exponential rise in the 



quantity of ribosomes observed in cells 
recovering from magnesium deficiency 
(where the quantity of DNA was high 
compared with the quantity of ribosomes), 
it appears that the rate of ribosome syn- 
thesis may depend both on the quantity 
of ribosomes and on the quantity of DNA. 

structure. Furthermore, it was noted that 
the ribosomes in elution from the column 
lost one-half their protein. The swinging- 
bucket technique has allowed these meas- 
urements to be made with separated 
groups of particles. 
Protein to nucleic acid ratio. A well 









.02 - 

Fig. 23. The exponential increase in ribosome content of E. coli recovering from magnesium 

Composition of the Ribosomes 

The composition of the ribosomes can 
give many clues to their structure and 
biosynthesis. In last year's report it was 
shown that ribosomes purified by repeated 
centrifugation and washing contained ap- 
proximately two amino acids per nucleo- 
tide, which suggested a three-stranded 

washed ribosome pellet derived from cells 
uniformly labeled with S 35 was suspended 
in 10" 4 M Mg to give 30 and 50S particles. 
These groups were separated in the swing- 
ing bucket and analyzed. The ratio of 
nucleic acid to protein (by Folin or S 35 ) 
was essentially the same in the two groups. 
The separated particles were then adsorbed 



to the DEAE column and eluted. The 
30S particles on elution in the A peak had 
half the original protein (by Folin or by 
S 35 ) per unit nucleic acid. A mixture of 
these particles with the original prepara- 
tion of 30S particles showed a single 
peak in the analytical centrifuge, indi- 
cating that the sedimentation constant was 
only slightly altered. Evidently the loss 
in mass is compensated by the increase 
in density. Similar results were obtained 
with the 50S particles. 

Fraction number 

Fig. 24. Swinging-bucket sedimentation anal- 
ysis of steady-state S 35 -labeled E. coli ribosomes. 
Optical density at 260 n indicates RNA content, 
radioactivity indicates protein content, and en- 
zyme activity indicates leucine aminopeptidase 
content. The particle peak at fraction 10 is com- 
prised principally of the 70 and 85S components. 

The constancy of the protein to nucleic 
acid ratios in the larger particles is shown 
in figure 24. The swinging-bucket tech- 
nique was used to separate the particles 
from cells uniformly labeled with S 35 . 
This method avoids the contamination of 
the ribosomes by precipitated protein which 
may be present in ribosome pellets. 

Cystine content of ribosomes. Experi- 
ments reported last year showed very little 
cystine in the ribosomes. This result had 
considerable significance as it indicated 

that ribosomal protein could not be the 
source of the soluble proteins which con- 
tain cystine. On repeating this measure- 
ment (using the mutant strain of E. coli 
ML 304D), cystine appeared in the ribo- 
somes. Another repetition with E. coli B 
again showed only traces of cystine. Fur- 
ther experiments then showed considerable 
cystine (cystine/methionine = 1 /q) in ribo- 
somes of E. coli B. As these measure- 
ments require only very simple procedures 
of hydrolysis and chromatography it is 
difficult to understand how they could fail 
to show any cystine present. On the other 
hand, it is equally difficult to believe that 
the cystine content of the ribosomes could 
be variable. 

Composition of the RNA. The nucleo- 
tide compositions of several classes of E. 
coli RNA have been determined by means 
of isotope dilution, using chromatographic 
and electrophoretic separations. The re- 
sults are presented in table 15. C, A, G, 
and U signify cytidylic, adenylic, guanylic, 
and uridylic acids, which are the principal 
building blocks of RNA. S is the RNA 
which is found in the supernatant fluid 
after a cell juice has been centrifuged (40K 
180'). This RNA is eluted on DEAE- 
cellulose in region B. CA-B is an RNA 
produced by cells which have been cul- 
tured for 2 l / 2 hours in a medium contain- 
ing 20 ug/ml of chloramphenicol. This 
RNA is also eluted from DEAE cellulose 
in region B. CA-D is RNA, produced in 
the presence of chloramphenicol, which is 
eluted in region D; 30(«) and 50(«) are 
the RNA's of the 30S and 50S particles 
which normally occur in juices of expo- 
nentially growing bacteria; 30(70) and 
50(70) are the RNA's of 30S and 50S 
particles which were derived from 70S 
particles by dissociation of the latter in a 
solution low in magnesium. It is evident 
from the results in table 15 that E. coli 
RNA may be classified according to com- 
position: one class, the "soluble" type, is 
characterized by a richness of cytidylic 
acid and a deficiency in adenylic acid; an- 
other, the "particle" type, is characterized 


by relatively more adenylic than cytidylic 
acid. It is also evident that the natural 
30S RNA has a composition very similar 
to that of the 30S material derived from 
the 70S particle. A similar result obtains 
for the 50S components. The RNA com- 
position of the 30S particle differs sig- 
nificantly from that of the 50S, however, 
the 50S particle being richer in purine 
nucleotides. Thus, it cannot be a simple 
dimer of 30S particles. The 70S particle, 
however, is a dissociable polymer, and it 
could be "synthesized" as a result of asso- 
ciation of natural 30S and 50S units. It 
would appear that the 30S and 50S are 

The DNA of E. coli contains nearly 
equimolar proportions of its component 
nucleotides, and the bases are paired in 
the arrangement G : C and A : T. Neither 
of these conditions is evidenced by the 
analyses of RNA. Thus, there is no simple 
correspondence between the composition 
of DNA and that of any of the RNA's 
examined. On the other hand, there is a 
high consistency in the proportions of 
6-amino : 6-keto groups, as shown by the 
ratio A + C: G + U in table 15. Thus, the 
RNA composition seems to be under some 
restraint. In addition, if it is assumed that 
all the pyrimidine residues are base-paired 

TABLE 15. Composition of E. coli Ribonucleic Acids 

Type of RNA, mole per 

cent * 





























* Values are the arithmetic means of several determinations for each component. The uncertainty 
is ± 3 per cent of the given value except for the bracketed value where the uncertainty was ± 7 
per cent. 

"fundamental" particles, the other species, 
70 to 100S, being derivatives. 

The results for the CA-RNA's are also 
of interest in the problem of nucleoprotein 
biosynthesis, since these are RNA's pro- 
duced in the virtual absence of protein 
synthesis. CA-B is clearly of the soluble 
type and is probably not associated with 
protein. The composition of CA-D, on the 
other hand, is like that of the nucleo- 
protein RNA. Thus, particle-type RNA 
can be made when protein synthesis is 
suppressed. In addition, the CA-D elutes 
in the region in which the polynucleotide 
precursor of nucleoprotein is found. CA-D 
has a small sedimentation coefficient (less 
than 20S), however, whereas the poly- 
nucleotide precursor sediments rapidly (20 
to 100S). 

(i.e., G:C and A:U), and the excess 
purine in the RNA's is computed, it is 
found that the 50S material has roughly 
twice the excess found for the 30S RNA. 
Again, there seems to be a kind of system- 
atics which hints at an underlying regu- 
lating mechanism. Many more data are 
needed, however, before more than vague 
generalizations can be expressed about the 
determinants of RNA composition. 

Enzyme Content of Ribosotnes 

Studies on the latent ribonuclease ac- 
tivity of E. coli ribosomes have been ex- 
tended, and another enzymic constituent 
of ribosomes, leucine aminopeptidase, has 
been identified. In addition, the quantity 
of 3:galactosidase associated with the ribo- 
somes has been measured. 


Ribonuclease. This activity is demon- 
strable when ribosomes are dissociated by 
treatment with 4.5 M urea or the chela- 
ter ethylenediaminetetraacetic acid. Upon 
treatment of ribosomes with either of these 
agents a rapid release of acid-soluble ultra- 
violet-absorbing components occurs. Most 
of the enzymic activity, as indicated by 
the digestion of added RNA, is found in 
the ribosomes of a cell juice, figure 25. The 
small amount (~10 per cent) found in 

Digestion of P32 lobeled E. coli 

RNA in 4.5M urea, pH 7.6 



/ Ribosomes 

/ * ^""* 


1 1 1 

15 30 45 

Minutes at 37° C 

Fig. 25. Ribonuclease activity of ribosomes 
and the soluble (SN) fraction of E. coli. The 
amount of enzyme is indicated by the slope of 
each line. 

the supernatant fluid after the ribosomes 
had been separated out is also latent. Since 
the supernates are rarely free of ribosomes 
it may be inferred that all the enzyme is 
particle-associated. After 10 minutes at 
37° C in 4.5 M urea essentially all the 
RNA of ribosomes is degraded, figure 26. 
The time for complete degradation is the 
same for mixtures of ribosomes, or puri- 
fied 70S, 50S, and 30S particles, and is the 
same whether the particles have been de- 
rived from cells growing in broth or in 

synthetic media, or from cells which were 
nongrowing as a result of having ex- 
hausted the glucose supply. Thus, a con- 
stant amount of enzyme activity in rela- 
tion to the total amount of nucleic acid 
in the ribosomes is observed. Since a 
given enzyme activity implies the exist- 
ence of unique molecules that have the 
capacity to act specifically, it may be in- 
ferred that there is a constant proportion 
of ribonuclease molecules in ribosomal par- 

Self-digestion of 
Ribosomes In 4.5 M 
ureo, pH 7.6 

10 20 

Minutes at 37° C 

Fig. 26. Digestion of ribosomal RNA by the 
latent ribonuclease in the ribosomes. Digestion 
is indicated by the appearance of acid-soluble 
ultraviolet-absorbing components. The results 
for three concentrations (1, 4, and 20 arbitrary 
units) are illustrated. 

tides. Attempts to determine this propor- 
tion accurately have been frustrating; our 
analyses indicate that there may be one 
ribonuclease molecule for each ribosomal 
particle as an upper limit, or that there 
may be ten ribosomal particles for each 
enzyme molecule as a lower limit. Until 
an accurate determination can be made, 
the important question whether ribonu- 
clease is a component of all ribosomes 
must remain open. The enzyme is a com- 


Acid soluble ° 5 
U. V. or P32 
from RNA 

.25 - 

40 60 

Minutes ot 37° C 

Fig. 27. The effect of urea concentration on the self-digestion of ribosomes. Urea concentrations 
of 2.4 to 4.5 M were used. 

ponent of some, at least, of the ribosomes, 
however, and it was of interest to inquire 
into the basis of its latent behavior. 

For this purpose ribosomes were sus- 
pended in various concentrations of urea 
at constant low ionic strength, or in solu- 
tions of 2.0 M urea with varying ionic 
strength, and the self-digestion of the in- 
herent ribosomal RNA was measured by 
the appearance of acid-soluble components. 

The results are shown in figures 27 and 

28. In the lower urea concentrations, or 
in the lower salt concentrations in 2.0 M 
urea, a considerable time lag is observed 
before digestion takes place. When diges- 
tion is initiated it appears to follow an 
autocatalytic course. 

Both urea and EDTA decompose ribo- 
somes, the former by breaking hydrogen 
bonds and the latter by chelating magne- 
sium. Thus, both hydrogen bonds and 
magnesium function to hold the ribosomes 



U. V 

20 30 40 

Minutes at 37 "C 

Fig. 28. The effect of NaCl concentration on the self-digestion of ribosomes in 2.0 M urea. The 
NaCl concentration varied from 0.0 to 0.4 M. 



together. The influence of NaCl in the 
decomposition of the ribosomes in 2.0 M 
urea (fig. 28) is most likely a result of 
the displacement of magnesium which has 
been made available to competition as a 
result of opening some hydrogen bonds. 
In the absence of urea, NaCl does not 
bring about ribosome self-digestion. The 
autocatalytic shape of the self-digestion 
curves thus appears to result from the 
loosening of the ribosomal structure until 
the enzyme is allowed to attack the RNA 
of a different particle and thereby release 
more enzymes. 

The RNAase of the ribosomes is not 
completely freed of the ribosomal struc- 
ture until digestion is complete. This was 
demonstrated by extracting the enzyme 
into acetate buffer after ribosomes had 
been precipitated with trichloroacetic acid 
over the course of an experiment in which 
self-digestion was allowed to take place. 
During the lag period (2.6 M urea, 37° C, 
TSM buffer) no enzyme was extracted. 
As digestion occurred ribonuclease activity 
appeared in the extract. Thus, the ribo- 
nuclease is held in the particle in the form 
of a TCA-precipitable nonacetate-extract- 
able complex until the particle is thor- 
oughly disrupted, at which time it be- 
haves as a "free" molecule. 

The interesting question of the intra- 
particle location of the enzyme was ex- 
amined in the following way: Nonradio- 
active ribosomes were treated to provide a 
lag period before self-digestion took place. 
Self -digestion was measured by the appear- 
ance of acid-soluble ultraviolet-absorbing 
material. At zero time P 32 -labeled RNA 
of high specific radioactivity was added. 
The digestion of the added substrate was 
measured by the appearance of P 32 in the 
acid-soluble fraction. The result is shown 
in figure 29. It is clear that the added 
RNA is largely digested before the in- 
herent RNA is attacked. Ribosomes that 
have not been urea-treated do not attack 
either substrate. Thus, the latency of some, 
at least, of the enzyme is very rapidly 
destroyed. It would appear that this la- 

tency is due to the existence of hydrogen 
bonds which keep the enzyme from act- 
ing. In addition, the ribonuclease of ribo- 
somes is so "located" that after partial 
degradation of the particle it is able to 
attack a high-molecular-weight exogenous 
substrate but it is not free as shown by 
the TCA-acetate procedure and is still 
unable to attack the RNA of the particle. 

Leucine aminopeptidase. Unlike ribo- 
nuclease activity the LAP activity of E. 
colt shows no latency. It appears fully 
active toward several leucyl substrates pro- 
vided that the magnesium concentration 
is adequate. The ribosomal association of 
this enzymic function was sought after it 
had been learned from Dr. A. T. Mathe- 
son, of McGill University, that LAP ac- 
tivity was found in a ribonucleoprotein 
fraction from autolyzed mammalian tis- 
sues. Crude ribosome preparations from 
E. coli invariably showed LAP activity 
against leucyl amino acid peptides. In 
some preparations the activity was as small 
as 10 per cent of the total in the cell, 
and in others as high as 80 per cent. 
The distribution between ribosomes and 
the soluble fraction apparently depended 
on the physiological state of the cells 
when they were harvested. For example, 
the ribosomes of resting cells which had 
been grown in a medium rich in glu- 
cose showed a greater fraction of the 
total LAP activities than those from cells 
grown in a low-glucose medium. The 
mere presence of enzyme activity in the 
soluble fraction raised the serious ques- 
tion whether the LAP activity of the 
ribosomes was an inherent capacity or one 
brought about as a result of adsorption. 

Several lines of evidence add up to show 
that some of the LAP activity of ribo- 
somes is indeed an inherent capacity. 
When the chromogenic substrate leucyl-3- 
naphthylamine was used in place of leu- 
cyltyrosine to carry out the assays, the 
ribosomes were found to be devoid of 
activity, all the enzymic function residing 
in the soluble fraction. Thus, at least two 
enzymes capable of splitting leucyl pep- 



tides existed. One was characteristically a 
soluble component. It seems unlikely that 
ribosomes would specifically exclude one 
LAP function and retain another, were 
adsorption responsible. A further test was 
made by analyzing the LAP activity of 
purified 30S and 50S particles. These 
were derived from an original 70 to 100S 
mixture and had been through five cycles 
of differential ultracentrifugation, being 
finally purified by the swinging-bucket 
sucrose-gradient method. The results of 
the analysis are given in table 16. 

fication procedure. In still another test 
DEAE-cellulose chromatography was em- 
ployed. Ribosomes and also the soluble 
fraction remaining were chromatographed 
and all the fractions analyzed for LAP 
activity against leucyltyrosine. The results 
are shown in figures 30 and 31. The 
soluble fraction contained at least three 
separable LAP activities. The ribosome 
chromatogram contained only one region 
which showed the function. The enzyme 
activity in this region corresponded in 
chromatographic behavior and also in sub- 

Acid soluble 
U. V. — — - 

.• u.v. 


Minutes at 37° C 

Fig. 29. Digestion of radioactive added RNA and of nonradioactive inherent ribosomal RNA 
in 2.6 M urea. The former is shown by the appearance of acid-soluble P 32 , and the latter by acid- 
soluble ultraviolet-absorbing components. 

TABLE 16. LAP Activity of 50S and 30S 






per Unit 





It may be observed from these results 
that the enzyme activity per unit protein 
is the same for the two particle types. 
Again, it seems unlikely that adsorption 
could account for such a distribution of 
enzyme activity and its persistence with 
the particles throughout the involved puri- 

strate specificity (table 17) to one of the 
soluble components. Thus, the ribosomes 
appear to contain only one of a possible 
three LAP activities. Centrifugal analysis 
of ribosomal particles by the swinging- 
bucket technique has shown that the en- 
zyme activity tracks upon the particle dis- 
tribution during sedimentation yielding 
essentially constant specific enzyme activi- 
ties across the pattern (fig. 24). Thus, the 
sedimentation constant for this bacterial 
LAP activity can be very large, of the 
order of 70 to 100S. This is in contrast to 
that for a mammalian LAP which has a 
sedimentation coefficient of approximately 
15S. The ribosomal LAP is dissociated 
from nucleoprotein by DEAE-cellulose 


chromatography, as figure 30 shows. The 
chromatographically separated LAP shows 
a small sedimentation constant (less than 
20S). The yield of LAP by this procedure 
is disappointingly small ('-'10 per cent of 
that in the ribosomes), most of the enzyme 
remaining irreversibly bound to the col- 
umn. In this respect ribosomal LAP dif- 

the assay for the function, the high resolv- 
ing power of the swinging-bucket method, 
and the tempting notion that soluble-type 
enzymes ought to be found on ribosomes, 
since the ribosomes were making them, 
prompted a re-examination of the prob- 
lem. The results are still tentative. It 
appears convincing, however, that a small 


30 40 

Fraction number 



Fig. 30. DEAE-cellulose chromatography of leucine aminopeptidases of the soluble fraction of 
E. coli. The protein content and eluting salt concentration are shown on the upper diagram, and 
the enzyme activity on the lower one. The activity against leucyltyrosine (LT) is indicated by the 
unshaded regions; that against leucyl-(3-naphthylamine (LNA) is shaded. 

fers from ribonuclease which chromato- 
graphs with the nucleoprotein (A peak) 
on the column. Thus, a different location 
in the ribosome for LAP and ribonuclease 
may be inferred. 

$-Galactosidase. The association of this 
enzyme with ribosomal particles has here- 
tofore been suspect. However, the extraor- 
dinary sensitivity and great specificity of 

amount of (3-galactosidase exists on ribo- 
somes in a latent form. The latent activity 
can be developed by the addition of spe- 
cific antisera, and an increase in enzyme 
activity is thus observed. 

When ribosomes from fully induced 
E. coli were separated in the swinging- 
bucket rotor and analyzed for $-gd\acX.o- 
sidase activity, a small amount of en- 



zyme could be detected in association with 
the particles. When these particles were 
treated with the corresponding antibody 
(derived from rabbits immunized to the 
3-galactosidase of E. coli or Bacillus me- 
gatherium) a several-fold increase in en- 
zyme activity was observed. Treatment 
of the soluble enzyme with the antisera 
showed no increase in activity. Normal 
rabbit serum did not increase either the 

minus bacteria, which are completely de- 
void of enzyme, were mixed with the 
enzyme in the soluble fraction of induced 
cells, separated by the swinging-bucket 
technique, and analyzed. Only a small 
amount of enzyme activity was detected. 
Unlike the others, however, the ribosomes 
from lac minus bacteria did not show an 
increase in enzyme activity when they 
were antibody-treated. Sixfold increases in 

10 20 30 40 50 

Fraction number 



Fig. 31. DEAE-cellulose chromatography of the leucine aminopeptidase of E. coli ribosomes. 
Protein and eluting salt concentration are shown on the upper diagram, and enzyme activity against 
leucyltyrosine (LT) is indicated on the lower one. 

particle or soluble enzyme activity. Anti- 
serum against cellulose-[/7-NH2-thio-3-/- 
galactoside] also had no effect. Generally 
similar results on the increase in particle- 
associated enzyme activity have been 
found for preparations from an inducible 
but noninduced strain and from a con- 
stitutive strain of E. coli. As a critical 
test to assay the possibility of an artifact 
due to adsorption as the explanation of 
these phenomena, the ribosomes of lac 

enzyme activity have been achieved with 
some preparations. It appears that some 
molecules of 3-galactosidase in E. coli are 
particle-associated and exist in a latent 
form. The particle enzyme can be acti- 
vated by appropriate antisera. 

Particles of Other Microorganisms 

A survey of the particle contents of rep- 
resentative microorganisms suggests that 
the results of the study of the ribosomal 


particles of E. coli may be of general 
applicability. Although the study was 
somewhat cursory, and no attempt was 
made to estimate accurate sedimentation 
constants, the general impression gained 
from ultracentrifuge analysis was one of 
remarkable uniformity. The organisms 
studied include six strains of bacteria, 
two yeasts, and one mold. The more 
rapidly growing bacteria appeared to have 
particle contents similar to those of E. coli, 
and at the other extreme Aspergillus niger 
mycelia have very few. 

In juices from growing cells particles of 
70 to 80S are always predominant. They 

TABLE 17. Specificity of Eluted Enzyme 



jle Enzyme 






































L-Leucylglycyl glycine 










amide hydrochlori 

de - 









are usually accompanied by much smaller 
quantities of one more rapidly and two 
more slowly sedimenting components cor- 
responding to the 100, 50, and 30S peaks 
of E. coli. There are, of course, differences 
in the proportions of the various particle 
sizes from one organism to another. In 
particular, juices from Pseudomonas fluo- 
rescens appear to contain relatively large 
amounts of the smallest particle (~20S). 
In all organisms studied the stability of 
the larger particles seems to be magne- 
sium-dependent, so that reduction of the 
Mg ++ ion concentration to 10" 4 M leaves 
only the 50 and 30S particles. 

Studies of other properties such as col- 
umn behavior and enzyme content are 

also of value in comparing particles of 
different origins. The behavior of par- 
ticles of E. coli, Aerobacter aerogenes, Ps. 
fluorescens, B. megatherium, and Salmon- 
ella typhimurium on columns of DEAE 
appears to be very similar. The fact that 
all these particles are eluted at the same 
salt concentration suggests that their struc- 
tures have much in common. 

Tests on highly purified particles from 
Ps. fluorescens, B. megatherium, and A. 
aerogenes showed the presence of both 
RNAase and peptidase activities. Self- 
digestion of the particle RNA by RNAase 
appeared to take place under the same 
conditions as described for E. coli, but no 
detailed comparison was attempted. The 
specificity of the peptidase reaction of 
these particles very closely resembles that 
of the E. coli ribosomes. Though there 
was evidence of RNAase activity in yeast 
ribosomes, no peptidase activities could be 

Synthetic Activities of the Ribosomes 

In growing bacteria, all the cellular com- 
ponents double in a generation time. These 
include the products of synthesis and also 
the synthetic machinery. Thus, when 
newly incorporated amino acids are found 
associated with ribosomes, a careful dis- 
tinction must be made as to whether these 
amino acids are destined to become soluble 
protein or whether they are structural 
components of newly formed ribosomes. 
Fortunately, there are a number of cri- 
teria by which the two processes can be 
distinguished, and it is possible to focus 
upon one or the other even though they 
are concurrent in the growing cell. 

Synthesis of nascent protein by ribo- 
somes. In Year Book 57 we reported that 
we had failed to find any precursors to the 
soluble proteins in association with the 
ribosomes. The newly incorporated amino 
acids found in the ribosomes fraction were, 
in fact, linked in the structural compo- 
nents of the ribosomes themselves. The 
literature contained several reports that 



the ribosomes of mammalian cells carried 
nascent protein. It, therefore, seemed quite 
strange that bacteria, which have a higher 
rate of protein synthesis and a higher con- 
tent of ribosomes, did not exhibit this 
phenomenon. It was clearly recognized, 
however, that the nascent protein might 
be associated with the ribosomes for only 
a very brief period and could thus be 
obscured by the accumulation of ribosomal 
protein. During the past year it was found 
possible to reduce the incorporation pe- 
riod drastically from minutes to seconds. 
These brief exposures to the tracers, to- 
gether with the improvements in sedimen- 
tation analysis, have given a clear demon- 
stration of the synthesis of the nascent 
protein 60 times more rapid than that 
previously observed in mammalian cells. 
It has also been possible to show the 
equally rapid transfer of the nascent pro- 
tein from its association with the ribo- 
somes to the soluble fraction. 

As a preliminary to the study of the 
incorporation into ribosomes, we investi- 
gated the kinetics of S 35 4 = fixation into 
the total cold TCA-precipitable fraction 
of E. coli and the rapidity with which 
this incorporation could be stopped by 
lowering the temperature or by adding a 
"chaser" containing nonradioactive SGr, 
cystine, and methionine. Samples were 
removed with a hypodermic syringe as 
rapidly as possible from a culture after 
addition of S 35 04, squirted into TCA, 
filtered, and counted. After an appropri- 
ate time the nonradioactive substrate was 
added and the sampling continued. As 
shown in figure 32 a linear rate of incor- 
poration was found after a few seconds, 
and this could be obliterated instantly 
when the "chaser" was added. These ex- 
periments were carried out with cells that 
had been grown for about 30 minutes 
without an exogenous source of sulfur. 
The absence of kinetic delays in the in- 
corporation of S 35 , and in its cessation, 
indicates that under these conditions of 
sulfur starvation there is little if any pool 

of precursors to the sulfur amino acids 
of the cell protein. 

In other experiments the culture was 
poured on ice a few seconds after the 
tracer was added. Samples taken from 
the iced suspension during the following 
2 hours showed a rate of incorporation 
less than 1/5000 of the rate at 37°. Fur- 
thermore, none of the S 35 that had been 
previously incorporated into TCA-precipi- 
table material was removed by exchange 

20 30 40 

Fig. 32. Time course of incorporation. S 35 4 = 
was added to a growing culture of sulfur-depleted 
E. coli at time 0. Samples were withdrawn at 
indicated times, and squirted into TCA, filtered, 
and counted. At 16 seconds S 32 4 = , S 32 cystine, 
and S 32 methionine were added. Note prompt 
incorporation of the tracer into TCA-precipi- 
table material and rapid cessation of incorpora- 
tion after addition of nonradioactive material. 

in the presence of chaser at the low 
temperature. Accordingly, changes in the 
content of TCA-precipitable S 35 occurred 
during the period when the cells were 
incubated at 37° and not during the period 
required for harvesting, washing, and frac- 
tionating the cells, all of which procedures 
were carried out between 0° and 4° C. 

Experiments of a few minutes' duration 
had consistently failed to show a higher 
specific radioactivity in the ribosomes than 
in the soluble protein. We therefore in- 
vestigated the distribution of S 35 after very 


short times of incorporation. Organisms 
were incubated for 10 seconds with S 35 C>4 = , 
and their components were separated by 
sedimentation analysis. The distribution 
of TCA-precipitable radioactivity and 260- 
mu adsorbing material is shown in figure 
33. Most of the S 35 is seen to be associated 
with the ribosomes of the 70 to 85S class 
and with the soluble protein which did 
not sediment under the conditions used. 
When the ribosomes were pelleted from 

duration (5 to 20 seconds) the major part 
of the label associated with ribosomes was 
found in the 70 to 85S fraction, and the 
specific radioactivity of these particles was 
always substantially greater than that of 
the 30 and 50S ribosomes. When incuba- 
tions lasted 20, 45, and 120 seconds and 
the resulting cell juices were analyzed, it 
was found that the specific radioactivity 
of the 70S particles had risen rapidly by 
20 seconds and thereafter increased only 



; - 


i :- 






i * 

~L \. 











i S r 





r i - 





F ' *\J? 


~L- J " 



1 1 1 1 1 1 1 1 



70S 50S 30S 

* i i 


18 24 30 

_l_l 1 I 1 I I I I : L 

4 8 12 16 20 

Fraction number 

Fig. 33. Sedimentation analysis of cell juice after 15 seconds' incorporation of S 35 4 = . S 35 4 = 
was added to a growing culture of sulfur-depleted E. coli. After 15 seconds the culture was chilled, 
harvested, washed, and broken. The cell walls and membranes were spun out and the juice was 
analyzed in the swinging-bucket rotor; (a) shows analysis of the total juice; (b) analysis of resus- 
pended ribosome pellet; (c) analysis of supernatant fluid after removal of ribosomes. The 85, 70, 
50, and 30S particles seen in the analytical centrifuge are partly resolved. Note lack of contamina- 
tion of ribosome region by soluble protein, (c). Centrifugation 75 minutes at 37,000 rpm. 

such a juice, and the resuspended pellet 
was similarly analyzed, most of the soluble 
protein fraction was eliminated but the 
70S particles still carried their original 
radioactivity (fig. 33). This indicates that 
the S 35 was relatively firmly bound to the 
ribosomes (see also below). When the 
supernatant fluid remaining after the ribo- 
somes had been pelleted was analyzed, it 
was found that very little of the radio- 
activity had moved down the centrifuge 
tube (fig. 33). 
In a variety of experiments of short 

a little up to 120 seconds. The specific 
radioactivity of the 50S ribosomes rose 
more slowly but surpassed that of the 
larger particles between 45 and 120 sec- 
onds. The soluble protein also became 
labeled rapidly, but because of the much 
greater amount had lower specific radio- 
activity at the earlier time. 

To demonstrate that one component of 
the cell acts as precursor to another it is 
necessary to show not only that the radio- 
activity of the suspected precursor rises 
rapidly when a tracer substrate is added 


but also that it is transferred to the prod- 
uct with equal rapidity when the radio- 
active substrate is replaced by a nonradio- 
active one. Cells were therefore incubated 
with the radioactive tracer for 10 to 15 
seconds and then a "chaser" of nonradio- 
active amino acids was added. After a 
short period all incorporation was ter- 
minated by pouring the culture on ice. 
Figure 34 shows that most of the radio- 
activity found in the ribosome fraction 
at the end of 15 seconds was subsequently 
transferred to the soluble protein during 

5 seconds or less and died away equally 
rapidly when the tracer was diluted out. 
(2) The saturation level was equivalent 
to the quantity of soluble protein synthe- 
sized in 3 seconds. (3) The decrease in 
the radioactivity of the ribosomes during 
the chase was roughly equal to the con- 
comitant increase of radioactivity in the 
soluble protein. 

To check whether the results obtained 
with S 35 incorporation were typical of 
other amino acids, similar experiments 
were carried out to observe the incorpora- 



T 500 

r L 

— ^ 

n r r 

1" 400 

- ji r- 


(M .300 

j L 



a 200 



r i- *i kJ 


""f 1 II 

12 16 20 

Fraction number 

Fig. 34. Sedimentation analysis of cell juice, (a) Cells incubated 15 seconds with S 35 4 = . (b) 
Cells incubated 15 seconds with S 35 4 = followed by 15 seconds' incubation with S 32 "chaser." (c) 
Fifteen seconds' incubation with S 35 4 = followed by 120 seconds with S 32 "chaser." Note transfer 
of radioactivity from 70-85S region to nonsedimentary region. Centrifugation 75 minutes at 37,000 

the 15 seconds of incubation with the 
"chaser." The removal was only slightly 
greater in 120 seconds. Even a 5-second 
"chase" was quite effective. Thus, there 
is a protein component which is transiently 
associated with the ribosomes and has all 
the characteristics that would be expected 
in a compulsory precursor of the soluble 
proteins. It appears that this nascent pro- 
tein is a polypeptide strand formed in the 
ribosome and subsequently released as 
soluble protein. Owing to the short periods 
involved it has not been possible to plot 
detailed time courses of these processes, 
but three features have been established. 
(1) The radioactivity of the 70 to 85S 
ribosomes built up to a saturation level in 

tion of a mixture of amino acids. (These 
experiments were carried out in coopera- 
tion with Dr. R. Hendler, who used a 
portion of the cultures to measure the 
kinetic behavior of lipid-bound amino 
acids.) Figure 35 shows that the incorpo- 
ration of C 14 from the mixture of C 14 
amino acids (obtained by hydrolyzing 
Chlorella protein) began promptly but 
could not be terminated rapidly by adding 
a large excess of C 12 amino acids. Ex- 
change of exogenous amino acids with 
the amino acids of the pool is not very 
rapid, and C 14 continued to enter the pro- 
tein fraction for roughly 20 seconds after 
the "chaser" was added. Consequently, 
a longer period was needed to show the 



transfer of the particle-bound nascent pro- 
tein to the soluble fraction. The results, 
however, were quite similar to those ob- 
tained with S 35 . After 5 and 12 seconds 
a large fraction of the C 14 incorporated 
into the TCA-precipitable fraction was 
associated with the 70 to 85S particles, as 
shown by the sedimentation analysis. Dur- 
ing the subsequent 120 seconds after the 
addition of C 12 amino acids the C 14 was 
transferred to the soluble fraction (fig. 36) . 
As a result of the longer times required 

cation that the labeled amino acids were 
in peptide linkage. There exist, however, 
complexes of soluble RNA or lipids with 
amino acids which are precipitated by cold 
TCA. A number of tests, carried out to 
establish the nature of the labeled material 
which is transiently associated with the 
ribosomes, demonstrated that: (1) It re- 
mained precipitable by TCA after solution 
in 1 N NaOH. (2) It was not extractable 
by hot alcohol after cold TCA precipita- 
tion. (3) It was not exchangeable by 


80 100 

Fig. 35. Time course of incorporation of C 14 amino acids. Experimental conditions similar to 
those of figure 32 except C 14 -labeled Chlorella protein hydrolysate was used as tracer. Note that 
incorporation continues after addition of C 12 amino acids. 

to dilute out the free amino acid pools 
there was more opportunity for incorpo- 
ration into the structural proteins of the 
ribosomes and the transfer was not quite 
as complete as could be observed when the 
S 3j tracer was used. Neither was it pos- 
sible to demonstrate the rapidity of the 
transfer from the particle to the soluble 
fraction. Nevertheless, the results show 
that the behavior of the S 35 amino acids 
is consistent with that of the other amino 
acids and can be used with confidence to 
study the processes of protein synthesis. 

In the experiments described above we 
have taken TCA-precipitability as an indi- 

incubation with an excess of nonradio- 
active amino acids. (4) It yielded a large 
variety of compounds (peptides) having 
different electrophoretic mobilities after 
partial hydrolysis by chymotrypsin or 12 
N HC1. These tests indicate that the bulk 
of the newly incorporated amino acids that 
were found associated with the ribosomes 
have the behavior to be expected of amino 
acids bound in the peptide linkage. 

The association of the nascent protein 
with the ribosomes, although transient in 
the growing cell, is quite stable in the dis- 
rupted cell juice. Preparations have shown 
the same specific activity in the ribosome 


fraction after 5 days at 4° C. The ribo- 
somes can be centrifuged down from the 
cell juice and the pellet shows the same 
specific radioactivity in the 70S region 
upon subsequent sedimentation analysis. 
We have not yet found conditions that 
cause the release of the nascent protein 
from the ribosomes. Incubation (15 min- 
utes at 37°) with adenosine triphosphate 
(ATP) or with amino acids in the pres- 
ence of cell juice did not release it. 

On the other hand, the soluble protein 
which was labeled with S 35 as early as 7 
seconds after addition of the tracer could 
not be distinguished from the bulk of the 
unlabeled soluble protein by chromatog- 
raphy on DEAE. There is no indication 
from column chromatography that any 
appreciable time is required (in growing 
cells) for conversion from nascent protein 
associated with the ribosomes to the final 
form of the polypeptide chain. 







n ■ 



h : 







1 1 1 1 1 1 1 1 

12 16 20 

Fraction number 

Fig. 36. Sedimentation analysis of cell juice, (a) Cells incubated 5 seconds with C 14 amino acids. 
(b) Twelve seconds' incubation, (c) Twelve seconds' incubation with C 14 amino acids followed 
by 120 seconds' incubation with C 12 amino acids. Note decrease in radioactivity of 70 to 85S region 
and increase of radioactivity in nonsedimentable region. Centrifugation 75 minutes at 37,000 rpm. 

Chromatography with diethylaminoethyl 
(DEAE) cellulose caused disintegration 
of the ribosomes yielding nucleoprotein 
which was eluted from the column but 
leaving about one-half of the ribosome 
protein and 95 per cent of nascent protein 
firmly bound. On splitting the ribosomes, 
by decreasing the magnesium concentra- 
tion of the medium, the nascent protein 
was found partly in the soluble proteins 
but mainly in the 50 and 30S particles 
which comprised the bulk of the nucleo- 
protein and in a few remaining 70S ribo- 
somes (fig. 37). 

A less detailed study was made of the 
incorporation of S 35 in the fraction of 
broken cells which sedimented rapidly. 
This fraction contained fragments of cell 
walls and membranes as well as some 
intact cells. Under steady-state conditions 
of labeling roughly 25 per cent of the 
total S 35 was in this fraction whereas 
organisms which had been incubated with 
tracer for only 10 to 15 seconds showed 
about 30 per cent. This finding might be 
taken to indicate the presence of a protein 
precursor. Sedimentation analysis showed 
that half of this material sedimented very 



rapidly (>1000S), however, as might be 
expected for cell walls and unbroken cells. 
A large part of the remainder is accounted 
for by contamination with ribosomes and 
soluble protein, leaving only a minor com- 
ponent which has the sedimentation prop- 
erties to be expected of small fragments of 
cell membrane. The specific radioactivity 
of this fraction was similar to that of the 
soluble protein. Furthermore, its radio- 
activity did not diminish after the addition 
of a "chaser." Thus, there is no evidence 
that this fraction contains a major protein 

are more directly involved in protein 

The results reported here indicate that 
a time scale of 5 seconds for the comple- 
tion of protein units is approximately cor- 
rect. Free amino acid pools might intro- 
duce some slight kinetic delays which 
would increase the apparent time, but 
there is no doubt that the nascent protein 
has a transient existence of less than 5 to 
10 seconds in the growing cell. 

It is not possible from these experiments 
to determine the size of the polypeptide 


_ (a) 

70S 50S 30S 




- r - 




- | 

"1 i 1 


- ; _ 

i.r" ! 



- | 





- r 



: f \i 








i i 

E 4 



16 20 


_ (b) 

70S 50S 30S 



I I I I I I I I I I I I I I I I I I I 

4 8 12 16 20 24 

Fraction number 

Fig. 37. Sedimentation analysis of ribosomes from cells exposed for 10 seconds to S 35 4= . (a) 
Ribosome pellet resuspended with buffer containing 0.1 0~ 2 M Mg to give mostly 70 to 85S ribosomes. 
(b) Pellet resuspended with 10~ 4 M Mg to give mostly 50 and 30S ribosomes. 

precursor. There is, however, a slight indi- 
cation that ribosomes associated with cell 
membranes may be more active in protein 
synthesis. When cells were lysed by treat- 
ment with lysozyme, followed by freezing 
and thawing, about half of the ribosomes 
were released and most of the remainder 
were detached from the residual mem- 
branous material by passing it through the 
pressure cell. It was found that the first 
fraction of ribosomes had only about half 
the specific radioactivity of the second. 
Possibly some of the particles exist free 
in the cell juice whereas others are more 
or less firmly bound to membranes and 

strands making up the nascent protein, as 
this calculation requires a knowledge of 
the fraction of the ribosomes that is active. 
The 30 and 50S ribosomes are certainly 
not the major loci of the nascent protein, 
but they are only minor components and 
it cannot be excluded that they carry a 
correspondingly small part of the nascent 
protein. Their role in forming nascent 
protein is obscured by the rapid labeling of 
their structural protein. 

It is possible that both the 70 and 85S 
particles synthesize nascent protein, but 
the 70S ones consistently show an initial 
higher specific radioactivity. Hence, it 


seems more likely that the nascent protein 
is formed on them and is carried on to 
the 85S particles by a rapid interchange 
between these forms. Finally, it is also 
possible that only the 70S particles associ- 
ated with the cell membrane are active. If 
we assume, however, that all the 70S par- 
ticles are equally active and that the others 
are inert, and if we take 5 seconds as the 
time for formation of a polypeptide strand, 
the product must have a molecular weight 
of roughly 20,000. 

Synthesis of the ribosomes. Studies of 
the synthesis of the ribosomes elucidate 
other processes in addition to the synthe- 
sis of nascent protein by the ribosomes. 
These are the synthesis of the protein and 
RNA of the ribosomes and the assembly 
or growth from small to large particles. 

To observe precursor-product relation- 
ships among the various sizes of ribosomes 
it is desirable to observe the distribution 
of radioactive tracers after short periods of 
incorporation. At very early times nascent 
protein carries most of the radioactivity. 
Accordingly, somewhat longer periods of 
incorporation (1 to 2 minutes) followed 
by 15 seconds of growth in nonradioactive 
medium are better suited for observation 
of synthesis of ribosomes. The longer pe- 
riod allows radioactivity to accumulate in 
the ribosomes whereas the nascent protein 
reaches its maximum level in a few sec- 
onds; the "chasing" period releases the 
radioactive nascent protein from the ribo- 
somes whereas the ribosomal protein is 
not affected. Using these precautions, con- 
tamination from nascent protein can be 

The incorporation into the particles after 
a brief exposure to S 35 C>4 = is indicated in 
figure 38. The radioactivity per unit of 
260-mu absorbing material (S*/UV) is 
highest in the small particles. Uniformly 
labeled cells show a constant S*/UV in 
the various sizes of particles. 

These data suggest that the small par- 
ticles are precursors of the larger ones, but 
the relationships are not simple. When 
cells are given a short period for incorpo- 

ration and then allowed to grow in non- 
radioactive medium, radioactivity persists 
in the small particles. The small particles 
cannot act solely as precursors but must 
exchange material with their products; 
otherwise they would soon lose their radio- 

Column chromatography also yields in- 
formation on the synthesis of the ribo- 
somes. It was shown above that ribosomes 
were severely degraded when they were 
adsorbed on and then eluted from DEAE 
columns. One-half the protein remained 
attached to the column, and accordingly 
one-half of the S 35 was lost from particles 
derived from uniformly labeled cells. The 
particles may be visualized as having two 
types of protein, one being a "coat" which 
is removed by the column and the other 
a "core" which remains associated with 
the nucleic acid. In contrast, particles 
labeled during a short exposure to S 35 O.T 
lose 90 per cent of their radioactivity on 
elution from the column. After 30 min- 
utes' additional incubation in nonradio- 
active medium the loss of radioactivity 
was reduced to 65 per cent. These char- 
acteristics are compared in table 18. One 
interpretation of these results is that the 
new protein of a growing particle is laid 
down as the "coat" on a "core" of pre- 
existing protein; subsequently, the "coat" 
acts as "core" for the next generation of 
particles. Alternatively, the newly formed 
particles may be more susceptible to deg- 

The soluble RNA appears to be unre- 
lated to the RNA of the ribosomes. The 
two types can be separated by sedimenta- 
tion, by column chromatography, or by 
electrophoresis. As mentioned above, the 
composition of the soluble RNA is differ- 
ent. Kinetic measurements showed that 
the soluble RNA was not a precursor to 
ribosomes. On the contrary, it was a 
stable end product of synthesis as its P 32 
content was undiminished after several 
generations' growth in a P :!1 medium. 

Neither was there any exchange of 
phosphorus between the soluble RNA and 



.600 — 



.100 — 

12 16 20 24 

Fraction number 

Fig. 38. Swinging-bucket analysis of ribosomes from E. coli labeled by uptake of S 35 for 1 minute. 

the microsomal RNA. After prolonged 
growth in a P 32 synthetic medium both 
the soluble RNA and the ribosomal RNA 
had the same specific radioactivity. The 
cells were then transferred to P 31 broth, 
which caused both the growth rate and 
the ratio of ribosomes to soluble RNA to 
increase. As a result there was relatively 
more new material in the ribosomes and 
their specific radioactivity was less than 
that of the soluble RNA. Had there been 
circulation or exchange between the two 
types they would have maintained equal 
specific radioactivities even though one 

increased in quantity relative to the other. 

Chloramphenicol, which blocked the 
formation of ribosomes, had no immediate 
effect on the synthesis of the soluble RNA. 
Cessation of DNA synthesis in 15T" cells 
had different effects on synthesis of soluble 
RNA and ribosomes. Thus, there seem to 
be many lines of evidence that indicate 
the disparity of these two materials. 

In last year's report it was shown that 
P 32 en route to the ribosomes passed 
through a transient form which could 
readily be distinguished from ribosomes 
by chromatography. Although this ma- 

TABLE 18. Comparison of Ribosomal and A Peak Nucleoproteins 

Type of 


Ribosome Pellet 



A Peak Eluted from 

AA/Base S*/UV 




8 hr S 35 4 = 
3 min S 35 4 
3 min S 3r '0 4 
30 min S 32 






terial was present in small quantities its 
properties could be determined readily as 
it contained the bulk of the P 32 incorpo- 
rated during a short incubation. It was 
sensitive to RNAase, it sedimented less 
rapidly than the 70 and 85S ribosomes, 
and it appeared in the D region of the 
elution pattern from DEAE. It had all 
the kinetic properties to be expected of a 
precursor to ribosomes. 

As ribosomes stripped of one-half their 
protein by DEAE chromatography also 
appear in the D region when rechromato- 
graphed it seemed quite possible that the 
precursor material consisted of incomplete 
ribosomes which had not yet received their 
full complement of protein. This view 
was strengthened by the observation that 
no newly incorporated S 35 was eluted in 
the D region together with the newly 
incorporated P 32 . 

This interpretation does not now seem 
tenable in view of further properties de- 
termined by electrophoresis and the im- 
proved technique of sedimentation anal- 
ysis. Electrophoresis can resolve soluble 
RNA and ribosomes as two separate peaks. 
The degraded ribosomes lacking one-half 
their protein move with an intermediate